JP7803182B2 - Image reading device and correction data creation method - Google Patents

Description

The present invention relates to an image reading device that performs shading correction and a method for creating correction data.

Image reading devices perform shading correction to reduce the effects of density variations between pixels caused by factors such as uneven illumination of the reading light source. A storage unit in the image reading device stores dark reference data, which is obtained when the image signal output from the image sensor when reading is performed with the image sensor's reading light source turned off and is converted into digital data, and white reference difference data, which is obtained when the image signal output from the image sensor when a pure white white reference image is read and is converted into digital data. The shading correction function based on the dark reference data and white reference difference data reduces the effects of density variations between pixels in the digital data obtained by converting the image sensor's output analog signal using an analog-to-digital (A/D) converter.
As an image reading device, a scanner is known that has a background plate provided opposite an image sensor. As described above, the whiter the background plate, the better for obtaining white reference difference data.

On the other hand, for documents with a white background, it is convenient to have a black background board for the purpose of finding missing parts of the document or for the image cropping function to eliminate white space in the scanned image. For this reason, some scanners are equipped with a function to switch the background board from white to black. Some documents may have dark areas such as photographs or drawings, and the blacker this black background board is, the more accurately it can be detected.

The image reading device disclosed in Patent Document 1 is equipped with a mechanism that moves and switches between a white background for shading correction and a black background for document extraction, allowing for switching between white and black backgrounds for reading.

JP 2017-98716 A

The image reading device described above has a complex structure because it is equipped with a mechanism for moving the background. Therefore, there is a demand for a background plate with a simple structure that can read both black and white backgrounds.

The image reading device of the present invention comprises:
a reading unit including a plurality of reading elements arranged along an arrangement direction intersecting a document conveyance direction, and configured to read an image of the document being conveyed;
an irradiation unit that irradiates a reading range by the reading unit with light;
a background plate facing the plurality of reading elements at a position where the light is irradiated;
a storage unit for storing data;
a control unit that controls the reading of the image,
the reading unit, the irradiation unit, and the background plate are arranged so that their positions in the conveying direction do not change;
the background plate includes a white background portion for shading correction and a black background portion for document cutout, the black background portion having a lower brightness than the white background portion;
the illumination unit includes a light source disposed on one side of the plurality of reading elements in the arrangement direction, and a light guide that guides the light from the light source to the reading range,
the white background portion is located farther from the light source than the black background portion in the arrangement direction,
The storage unit
background board reference difference data representing the difference between the read data of the background board irradiated with the light and the read data of the background board not irradiated with the light, assuming that the black background portion is the white background portion;
white reference difference data representing a difference between read data of the white reference chart irradiated with the light in a state where the white reference chart is placed in a range where the image is read and read data of the background board not irradiated with the light,
The control unit
acquiring black reference read data, which is read data of the background board to which the light is not irradiated, from the reading unit;
background plate read data, which is read data of the background plate irradiated with the light, is acquired from the reading unit;
From the background board read data, a black background read value at a specific position in the black background portion that is closer to the end of the white background portion side than the end of the light source side among both ends in the arrangement direction, and a plurality of white background read values at a plurality of read positions in the white background portion in the arrangement direction are obtained;
determining a virtual white background read value that is assumed as a read value at the specific position when it is assumed that the specific position is the white background portion based on the plurality of white background read values;
converting the background board read data into virtual background board read data corresponding to the read data of the white background portion based on a difference between the black background read value and the virtual white background read value;
The correction data for shading correction is calculated based on the difference between the virtual background board read data and the black reference read data, the background board reference difference data, and the white reference difference data.

The correction data creation method of the present invention further comprises:
a reading unit including a plurality of reading elements arranged along an arrangement direction intersecting a document conveyance direction, and configured to read an image of the document being conveyed;
an irradiation unit that irradiates a reading range by the reading unit with light;
a background plate facing the plurality of reading elements at a position where the light is irradiated;
a storage unit that stores data,
the reading unit, the irradiation unit, and the background plate are arranged so that their positions in the conveying direction do not change;
the background plate includes a white background portion for shading correction and a black background portion for document cutout, the black background portion having a lower brightness than the white background portion;
the illumination unit includes a light source disposed on one side of the plurality of reading elements in the arrangement direction, and a light guide that guides the light from the light source to the reading range,
the white background portion is located farther from the light source than the black background portion in the arrangement direction,
The storage unit
background board reference difference data representing the difference between the read data of the background board irradiated with the light and the read data of the background board not irradiated with the light, assuming that the black background portion is the white background portion;
a white reference difference data representing a difference between read data of the white reference chart irradiated with light in a state where the white reference chart is placed in a range where the image is read and read data of the background board not irradiated with the light, the white reference difference data including:
a first step of acquiring black reference read data, which is read data of the background plate to which the light is not irradiated, from the reading unit;
a second step of acquiring background plate read data from the reading unit, the background plate being read by the light;
a third step of acquiring, from the background board read data, a black background read value at a specific position in the black background portion that is closer to the end of the white background portion side than the end of the black background portion that is closer to the light source side, and a plurality of white background read values at a plurality of read positions in the white background portion in the arrangement direction;
a fourth step of calculating a virtual white background read value, which is assumed to be a read value at the specific position when it is assumed that the specific position is the white background portion, based on the plurality of white background read values;
a fifth step of converting the background board read data into virtual background board read data corresponding to the read data of the white background portion based on a difference between the black background read value and the virtual white background read value;
and a sixth step of determining correction data for shading correction based on the difference between the virtual background board read data and the black reference read data, the background board reference difference data, and the white reference difference data.

FIG. 1 is a schematic cross-sectional view showing a configuration of an image reading device. FIG. 2 is a schematic plan view showing the configuration of a transport path in the image reading device. 1 is a schematic block diagram showing the overall configuration of an image reading device; FIG. 2 is a schematic diagram showing a background plate and an image sensor. 10 is a flowchart of a pre-processing in a reference example. FIG. 2 is a diagram illustrating output characteristics of an image sensor. 10 is a flowchart for calculating white shading data in a reference example. 10 is a flowchart of a pre-processing. FIG. 10 is a diagram showing the output of a reading unit corresponding to a background plate. 10 is a flowchart for calculating white shading data. FIG. 10 is a diagram showing a first pattern of document cutting processing. FIG. 10 is a diagram illustrating a second pattern of document cutting processing. FIG. 10 is a diagram illustrating a third pattern of document cutting processing. FIG. 10 is a schematic diagram showing a background plate and a reading unit according to another example. FIG. 10 is a diagram illustrating a fourth pattern of document cutting processing. FIG. 10 is a diagram illustrating a fifth pattern of document cutting processing. FIG. 10 is a schematic diagram showing a background plate and a reading unit according to another example. FIG. 10 is a diagram illustrating a sixth pattern of document cutting processing. FIG. 10 is a schematic diagram showing a background plate and a reading unit according to another example. FIG. 10 is a schematic diagram showing a background plate and a reading unit according to another example. FIG. 10 is a diagram illustrating a sixth pattern of document cutting processing. FIG. 10 is a schematic diagram showing a background plate and a reading unit according to another example. FIG. 2 is a diagram schematically illustrating an example of the configuration of a reading unit. FIG. 2 is a diagram schematically illustrating a configuration example of an irradiation unit. 10A and 10B are diagrams schematically illustrating an example of the positional relationship between a plurality of reading elements arranged in the arrangement direction and a background plate in which the white background portion is located farther from the light source than the black background portion in the arrangement direction. 10A and 10B are diagrams illustrating an example of the relationship between the position of the reading element in the alignment direction and the read value. 10 is a diagram schematically illustrating an example of converting background plate read data into virtual background plate read data corresponding to read data of a white background portion. FIG. 10 is a diagram schematically showing an example of converting background board read data into virtual background board read data based on the difference between a black background read value and a virtual white background read value. FIG. 10 is a flowchart schematically illustrating an example of a pre-processing for calculating background plate reference difference data and white reference difference data and storing them in a storage unit. 10 is a flowchart schematically illustrating a white shading correction process. 10A and 10B are diagrams schematically illustrating an example of output characteristics relative to the position of a reading element in the alignment direction. FIG. 2 is a diagram schematically illustrating an example of output characteristics of a sensor chip.

Embodiments of the present invention are described below. Of course, the following embodiments are merely examples of the present invention, and not all of the features shown in the embodiments are necessarily essential to the solution of the invention.

(1) Overview of the technology included in this invention:
First, an overview of the technology included in the present invention will be described with reference to the examples shown in Figures 1 to 32. Note that the figures in this application are diagrams showing schematic examples, and the magnification in each direction shown in these figures may differ, and the figures may not be consistent with each other. Of course, each element of the present technology is not limited to the specific example indicated by the symbol. In the "Outline of the Technology Included in the Present Invention," the words in parentheses indicate supplementary explanations for the immediately preceding words.

[Aspect 1]
An image reading device 10 according to one aspect of the present technology includes a reading unit (e.g., an image sensor 13e) including multiple reading elements 30 arranged along an arrangement direction D2 intersecting a conveyance direction D1 of the document G, and reading an image of the document G as it is conveyed. The reading unit (13e) includes an irradiation unit 20 that irradiates a reading range AR1 of the reading unit (13e) with light LT1. A background plate 13c facing the multiple reading elements 30 at a position where the light LT1 is irradiated is also included. A storage unit (e.g., a non-volatile memory 17) that stores information is included. A control unit 15 that controls the reading of the image is also included. The reading unit (13e), the irradiation unit 20, and the background plate 13c are arranged so that their positions in the conveyance direction D1 do not change. As illustrated in FIG. 25 and other figures, the background plate 13c includes a white background portion 13c1 for shading correction and a black background portion 13c2 for document cutout that is lower in brightness than the white background portion 13c1. 24, the irradiation unit 20 includes a light source (e.g., LED 13d) disposed on one side of the plurality of reading elements 30 in the arrangement direction D2, and a light guide 23 that guides the light LT1 from the light source (13d) to the reading range AR1. As illustrated in Fig. 25 etc., the white background portion 13c1 is located farther from the light source (13d) than the black background portion 13c2 in the arrangement direction D2. The memory unit (17) stores background board reference difference data (e.g., background board RAW data GR'-BK) that represents the difference between the read data of the background board 13c irradiated with the light LT1 (e.g., virtual background board read data GR') and the read data of the background board 13c that is not irradiated with the light LT1 (e.g., black reference read data BK) when it is assumed that the black background portion 13c2 is the white background portion 13c1, and white reference difference data (e.g., white reference RAW data WH-BK) that represents the difference between the read data of the white reference chart irradiated with the light LT1 (e.g., white reference read data WH) and the read data of the background board 13c that is not irradiated with the light LT1 when a white reference chart is placed in the range for reading the image (e.g., maximum effective pixel width W1).
As illustrated in Figure 30, the control unit 15 acquires black reference read data BK_S, which is read data of the background board 13c when the light LT1 is not irradiated, from the reading unit (13e). The control unit 15 acquires background board read data GR_S, which is read data of the background board 13c when the light LT1 is irradiated, from the reading unit (13e). As illustrated in Figures 28 and 30, the control unit 15 acquires, from the background board read data GR_S, a black background read value Lb at a specific position (e.g., pixel n0) in the black background portion 13c2 that is closer to the white background portion side end 13c4 than the light source side end 13c3, and multiple white background read values Lw1, Lw2 at multiple read positions (e.g., pixels n1, n2) in the white background portion 13c1 in the alignment direction D2. The control unit 15 determines a virtual white background read value L0, which is assumed to be the read value at the specific position (n0) when it is assumed that the specific position (n0) is the white background portion 13c1, based on the multiple white background read values Lw1 and Lw2. The control unit 15 converts the background board read data GR_S into virtual background board read data GR_S' corresponding to the read data of the white background portion 13c1, based on the difference between the black background read value Lb and the virtual white background read value L0. The control unit 15 determines correction data for shading correction (e.g., white shading data SD) based on the difference between the virtual background board read data GR_S' and the black reference read data BK_S, the background board reference difference data (GR'-BK), and the white reference difference data (WH-BK).

In the above embodiment, virtual background board read data GR_S' is obtained from read data (GR_S) of the background board 13c, which includes the black background portion 13c2 and the white background portion 13c1, assuming that the black background portion 13c2 is the white background portion 13c1, and correction data (SD) for shading correction is obtained. Therefore, the above embodiment can provide an image reading device that does not require a mechanism for switching between a white background for shading correction and a black background for document extraction. As a result, the number of parts in the image reading device can be reduced.

Here, the "white" of the white background portion and the "black" of the black background portion are terms used for convenience to indicate high and low brightness, and are not limited to pure white or pure black. Therefore, the white background portion for shading correction only needs to be brighter than the black background portion for document cutout, and is not limited to pure white and can be gray. The black background portion for document cutout only needs to be brighter than the white background portion for shading correction, and is not limited to pure black.
The background board reference difference data is not limited to the difference between the reading data of the background board illuminated with light and the reading data of the background board not illuminated with light, but may also be a combination of the "reading data of the background board illuminated with light" itself and the "reading data of the background board not illuminated with light" itself.
The white reference difference data is not limited to the difference between the read data of the white reference chart illuminated with light and the read data of the background board not illuminated with light, but may be a combination of the "read data of the white reference chart illuminated with light" itself and the "read data of the background board not illuminated with light" itself.
The number of reading positions in the white background portion may be two, or may be three or more.
The above remarks also apply to the following aspects.

[Aspect 2]
28, the control unit 15 may obtain the virtual white background read value L0 at the specific position (n0) by performing an approximation calculation (for example, linear approximation) from the relationship between the plurality of white background read values Lw1, Lw2 for the plurality of reading positions (n1, n2) in the arrangement direction D2. This aspect obtains the virtual white background read value L0 taking into consideration the positional relationship between the light source (13d) and the reading position, thereby making it possible to provide an image reading device that can perform shading correction with even greater accuracy.

[Aspect 3]
The control unit 15 may perform a process to obtain the correction data (SD) at least one of before reading the document G and when power is turned on to the image reading device 10. This aspect can provide an image reading device suitable for performing shading correction.

[Aspect 4]
25 , the plurality of reading elements 30 may include a plurality of first elements 31 that are present in a maximum range (e.g., a maximum effective pixel width W1) for reading the image in the arrangement direction D2, and a plurality of second elements 32 that are present outside the maximum range (W1) in the arrangement direction D2. The plurality of first elements 31 may be present within a range AR2 of the black background portion 13c2 in the arrangement direction D2.
In the above case, the maximum range (W1) for reading an image in the alignment direction D2 is within the range AR2 of the black background portion 13c2, so the image can be extracted with high precision. Also, since the white background portion 13c1 exists outside the maximum range (W1) for reading an image in the alignment direction D2, it is possible to prevent the image reading device from becoming too large in the alignment direction.

[Aspect 5]
25, 28, etc., the specific position (n0) may be the position of the boundary B1 of the maximum range (W1) in the alignment direction D2. In this embodiment, the virtual white background read value L0 can be obtained at the position of the boundary B1 that is close to the reading position of the white background portion 13c1 within the maximum range (W1) for reading an image in the alignment direction D2, and therefore it is possible to provide an image reading device that can perform shading correction with even greater accuracy.

[Aspect 6]
25, 28, etc., the range of the plurality of reading elements 30 facing the background plate 13c in the arrangement direction D2 may be narrower than the combined range of the black background portion 13c2 and the white background portion 13c1 (e.g., ranges AR2 and AR3). In the arrangement direction D2, the range of the plurality of first elements 31 facing the black background portion 13c2 may be narrower than the range AR2 of the black background portion 13c2. In the arrangement direction D2, the range of the plurality of second elements 32 facing the background plate 13c may be a range that includes at least a portion of the white background portion 13c1. This aspect can provide a suitable example of an image reading device that eliminates the need for a mechanism for switching between a white background for shading correction and a black background for document extraction.

[Aspect 7]
Incidentally, as illustrated in FIG. 30 , a correction data creation method according to one aspect of the present technology is a correction data creation method for an image reading device 10 including the reading unit (13e), the irradiation unit 20, the background plate 13c, and the memory unit (17), and includes the following steps (A1) to (A6).
(A1) A first step (for example, step S402) of acquiring black reference read data BK_S, which is read data of the background plate 13c on which the light LT1 is not irradiated, from the reading unit (13e).
(A2) A second step (for example, step S404) of acquiring background board read data GR_S, which is read data of the background board 13c irradiated with the light LT1, from the reading unit (13e).
(A3) A third step (e.g., step S406) of acquiring from the background board reading data GR_S a black background reading value Lb at a specific position (n0) in the black background portion 13c2 that is closer to the end 13c4 on the white background portion side than the end 13c3 on the light source side of both ends in the alignment direction D2, and multiple white background reading values Lw1, Lw2 at multiple reading positions (n1, n2) in the white background portion 13c1 in the alignment direction D2.
(A4) A fourth step (e.g., step S408) of determining a virtual white background reading value L0, which is assumed to be the reading value at the specific position (n0) when it is assumed that the specific position (n0) is the white background portion 13c1, based on the multiple white background reading values Lw1 and Lw2.
(A5) A fifth step (e.g., step S410) of converting the background board read data GR_S into virtual background board read data GR_S' corresponding to the read data of the white background portion 13c1 based on the difference between the black background read value Lb and the virtual white background read value L0.
(A6) A sixth step (e.g., steps S412 to S414) of determining correction data (SD) for shading correction based on the difference between the virtual background board read data GR_S' and the black reference read data BK_S, the background board reference difference data (GR'-BK), and the white reference difference data (WH-BK).

The above aspect provides a correction data creation method that does not require a mechanism for switching between a white background for shading correction and a black background for document extraction.

Furthermore, the present technology can be applied to an image reading system including the above-mentioned image reading device, an image reading method including the above-mentioned correction data creation method, a control program for the above-mentioned image reading device, a computer-readable medium on which the control program is recorded, etc. The above-mentioned image reading device may be composed of multiple distributed parts.

(2) Description of Related Art:
First, techniques related to the present technology will be described with reference to Figures 1 to 22. Here, the "black reference data" shown in Figures 5 to 10 corresponds to the above-mentioned "black reference read data," the "background board data" shown in Figures 5 to 10 corresponds to the above-mentioned "background board read data," and the "white reference data" shown in Figures 5 to 10 corresponds to the above-mentioned "white reference read data."

Figure 1 is a schematic cross-sectional view showing the configuration of the image reading device, Figure 2 is a schematic plan view showing the configuration of the transport path CO1 of the image reading device, and Figure 3 is a schematic block diagram showing the overall configuration of the image reading device. In the X-Y-Z coordinate system shown in Figures 1, 23, 24, etc., the X direction is the width direction of the image reading device 10, which is also the width direction of the original G and the arrangement direction D2 of the multiple reading elements 30. The Y direction is a direction perpendicular to the X direction and is also the transport direction D1 of the original G. The Z direction is a direction perpendicular to the X and Y directions and is also a direction roughly perpendicular to the surface of the original G. Here, the X and Y directions do not have to be perpendicular as long as they intersect each other, the X and Z directions do not have to be perpendicular as long as they intersect each other, and the Y and Z directions do not have to be perpendicular as long as they intersect each other.

The image reading device 10 includes a mounting section 11 on which a document G is placed, a transport section 12 that transports the document G placed on the mounting section 11 along a predetermined transport path CO1, a reading unit 13 that reads the transported document G, a detection section 14 that detects the presence or absence of the document G along the transport path CO1, a control section 15 that controls the image reading device 10, a touch panel 16 that allows display and input, and a non-volatile memory 17 that stores data. Here, the non-volatile memory 17 is an example of a storage section. As illustrated in FIG. 3, the reading unit 13 includes a background plate 13c, an illumination section 20 having an LED (light-emitting diode) 13d, and an image sensor 13e. As illustrated in FIGS. 3 and 4, the image sensor 13e includes multiple sensor chips 13f (see FIG. 4) arranged along the alignment direction D2, each of which includes multiple reading elements 30 (see FIG. 3). The control unit 15 includes a processor such as a CPU (Central Processing Unit), ROM (Read Only Memory), and RAM (Random Access Memory). The CPU reads information stored in the non-volatile memory 17 into the RAM as needed, and executes the read programs to perform various processes. In a system in which the image reading device 10 is connected to a control device such as an external personal computer, the control device may also perform control and judgment together with the control unit 15.

The conveying section 12 includes a feed roller 12a arranged below the conveying path, a separation roller 12b arranged above the conveying path opposite the feed roller 12a, conveying rollers 12c arranged above and below the conveying path, and discharge rollers 12d arranged above and below the conveying path.
A transport roller 12c and a discharge roller 12d, which are disposed below the transport path, are connected to a drive mechanism and are driven to rotate. The transport roller 12c transports the document G to the reading unit 13, and the discharge roller 12d transports and discharges the document G that has been read by the reading unit 13.

The feed roller 12a and the separation roller 12b are disposed in the center of the conveyance path in the width direction. The conveyance roller 12c and the discharge roller 12d are disposed downstream of the feed roller 12a and the separation roller 12b in the conveyance direction. The width of each of the conveyance roller 12c and the discharge roller 12d is wider than the width of the feed roller 12a. The width of each of the conveyance roller 12c and the discharge roller 12d is also wider than the width of the separation roller 12b.
The detection unit 14 detects the presence or absence of an original G at each point on the transport path, and is equipped with a loading section detection sensor 14a that detects an original at the loading section 11, an original detection sensor 14b that detects an original at a downstream position near the transport roller 12c, and a double feed detection sensor 14c that detects double feeds.

The reading unit 13 comprises a pair of reading units 13a, 13b facing each other across the transport path. Each reading unit 13a, 13b comprises a background plate 13c, an LED 13d as a light source, and an image sensor 13e as a reading section. The LED 13d emits light toward the facing document, and the image sensor 13e outputs a detection signal according to the intensity of the reflected light. The background plate 13c is located on the opposite side of the document. The image sensor 13e has multiple reading elements arranged in the width direction of the transport path, with each reading element corresponding to one pixel. The reading units 13a, 13b read images on both sides of the document G as it is transported along the transport direction. The image sensor is also called a line sensor.

The image sensor 13e of the reading unit 13 is positioned so that it intersects with the document transport path, and corresponds to a reading section that reads the image of the document G passing through the transport path. The background plate 13c is positioned opposite the image sensor 13e, which is the reading section, across the transport path. In this example, the background plate 13c and image sensor 13e are fixed in predetermined positions so that they do not move in the document transport direction. Furthermore, because they are fixed, the background plate 13c and image sensor 13e do not rotate.

The touch panel 16 displays a predetermined image based on instructions from the control unit 15, and when the user touches it, outputs the touch position to the control unit 15. The control unit 15 can instruct the user to display a menu, or obtain the user's operation content based on the user's touch position.

FIG. 4 is a schematic diagram showing a background plate and an image sensor.
Image sensor 13e has many reading elements arranged in the width direction of the transport path, but in this example, the remaining reading elements are used, excluding the reading elements at both ends and their vicinity, and the width of the reading elements used is the maximum effective pixel width. Background board 13c is arranged opposite the reading area of image sensor 13e, and both ends of the transport path in the width direction are white background areas 13c1, 13c1 with high brightness for shading correction, and the area between them is black background area 13c2 with lower brightness than white background area 13c1 for shading correction and document cutout. In other words, white background area 13c1, black background area 13c2, and white background area 13c1 are arranged side by side in the main scanning direction of reading by image sensor 13e, which is the reading unit.

In this way, the white background portion 13c1 is provided in the areas at both ends of the background plate 13c in the direction intersecting the transport path, and the black background portion 13c2 is provided in the area between the white background portions 13c1, 13c1.

The white background portion 13c1, which has a high brightness, generally refers to a color that is as white as possible, and the black background portion 13c2, which has a lower brightness than the white background portion 13c1, generally refers to a color that is as black as possible. There is no specific brightness threshold. Also, the white background portion 13c1 can be substituted for a medium gray.

The black background portion 13c2 located in the center is wider than the width of the expected main document G, and in comparison, the white background portions 13c1 located at both ends only occupy the remaining area at both ends. In other words, the area of the black background portion 13c2 is larger than the area of the white background portion 13c1. For example, if the expected main document G is located exactly in the center of the width of the transport path and is transported straight without distortion, parts of the black background portion 13c2 will appear on both sides of the width of the document G, and the image sensor 13e will be able to read the black background portion 13c2 as an image.

The white background portion 13c1 and the black background portion 13c2 of the background plate 13c are arranged on the same plane and are the same distance from the image sensor 13e, which is the reading unit. The image sensor 13e, which is the reading unit, has multiple reading elements arranged in the main scanning direction, with a portion of the white background portion 13c1 positioned opposite an effective reading element near the end of the reading element. In this example, the reading element faces the white background portion 13c1 at both ends within the range of the maximum effective pixel width, but it may also face only the white background portion 13c1 at one end within the range of the maximum effective pixel width.
By making the distance from the image sensor 13e, which is the reading unit, to the background plate 13c the same, the accuracy of shading correction and extraction processing is improved.

In the image reading device 10 of this example, the starting end of the transport path is provided with a loading section 11 and feed roller 12a, which serve as a document supply section for supplying documents G. Furthermore, approximately in the center of the width of the transport path, a separation roller 12b is provided as a separation means to prevent multiple documents from being fed. In terms of positional relationship with this separation means, the black background portion 13c2 in this example is located in a position that includes the separation roller 12b, which serves as a separation means, in the width direction, which is the direction intersecting the transport direction, while the white background portions 13c1, 13c1 are located in positions that do not include the separation roller 12b, which serves as a separation means.

In addition, the document detection sensor 14b is located approximately in the center of the transport path, the black background portion 13c2 is located in a position that includes the document detection sensor 14b in the width direction, which is the direction that intersects with the transport direction, and the white background portions 13c1, 13c1 are located in positions that do not include the document detection sensor 14b.
Similarly, the double feed detection sensor 14c is located near the middle of the transport path, although not exactly in the center, and the black background portion 13c2 is located in a position that includes the double feed detection sensor 14c in the width direction based on the transport direction, while the white background portions 13c1, 13c1 are located in positions that do not include the double feed detection sensor 14c. Also, between the white background portions 13c1, 13c1 are the separation roller 12b, the transport roller 12c, the discharge roller 12d, the document detection sensor 14b, and the double feed detection sensor 14c.
The white background portions 13c1, 13c1 and the black background portion 13c2 are located opposite the LED 13d.

FIG. 5 is a flowchart of the pre-processing of the reference example, FIG. 6 is a diagram showing the output characteristics of the image sensor, and FIG. 7 is a flowchart of the calculation of white shading data of the reference example.
The input-to-output relationship of each reading element that makes up the image sensor 13e is shown in Figure 6. In this figure, the horizontal axis represents input and the vertical axis represents output. When a black background (black target) is imaged without illumination on the reading element, and when an image of a reference chart (white target) that is as white as possible is imaged with illumination, the output value ranges from the darkest input to the brightest input, as shown on the horizontal axis. This shows that the output of the reading element varies within the range of 0 to 255.

As shown in Figure 6, the input changes within the range from actual black (minimum) to white (maximum) and does not necessarily change from zero to the maximum value as an absolute value, so the output also does not change from zero (0) to the maximum value (255). For this reason, a process is required to associate the change range of the output element of each reading element with the change range of the actual maximum change range of light and dark.
This correspondence requires processing at the time of shipment from the factory and processing at the start of scanning. Figure 5 shows the processing at the time of shipment from the factory. As mentioned above, the background plate in the reference example can be switched between white or gray and black, and the white or gray side is set as the background plate side at the time of shipment from the factory below.

In step S102, the LED is turned off and the image sensor is caused to perform a reading operation to obtain black reference data BK; in step S104, the LED is turned on and the image sensor is caused to perform a reading operation of the background board to obtain background board data GR; in step S106, the background board RAW data is calculated based on the following equation (1), and then the background board RAW data is stored in non-volatile memory.
Background board RAW data = GR - BK ... (1)

Next, in step S108, with the LED turned on, the white reference chart is placed on the transport path, and the image sensor is caused to perform a reading operation to acquire white reference data WH. In step S110, the white reference RAW data is calculated based on the following formula, and then the white reference RAW data is stored in non-volatile memory.
White reference RAW data = WH - BK ... (2)
In this way, the background plate RAW data and white reference RAW data are stored in the nonvolatile memory at the time of shipment from the factory.

The input/output characteristics of the image sensor change due to various factors, including aging, so a process to calculate white shading data to correct this is performed immediately before scanning. In this case, too, a white or gray surface is used as the background board surface.

In step S122, the image sensor is made to perform a reading operation with the LED turned off to obtain black reference data BK_S, and in step S124, the image sensor is made to perform a reading operation of the background board with the LED turned on to obtain background board data GR_S, and in step S126, the previous background board RAW data is calculated based on the following equation (3).
Previous background board RAW data = GR_S - BK_S ... (3)

Next, in step S128, the white reference RAW data and background board RAW data stored in the nonvolatile memory, and the calculated previous background board RAW data are used to calculate white shading data based on the following equation (4).
White shading data =
(White reference RAW data) / (Background board RAW data) x Previous background board RAW data ... (4)

The value Dout obtained by performing shading correction on the input pixel data Din is expressed by the following equation (5).
Dout=(Din-Bk)/(Wh-Bk)×255…(5)
Once the white shading data has been calculated, the surface of the background board is mechanically changed to a black surface. This is because a black background board increases the difference in brightness between it and the original, which improves the accuracy of original extraction.

In order to calculate white shading data using this type of processing while still trying to improve the accuracy of document extraction, the reference example had no choice but to provide a background plate with two sides, white and black, and a mechanism for switching between the sides.
In contrast, the background plate 13c of this embodiment is not configured to switch sides, as shown in FIG. 4, but has white background portions 13c1 on both ends and a black background portion 13c2 in the center.

For this reason, the reading element of the image sensor 13e facing the white background portion 13c1 cannot measure the input/output relationship when facing the black background portion, and the reading element of the image sensor 13e facing the black background portion 13c2 cannot measure the input/output relationship when facing the white background portion.
In this example, in consideration of situations where measurements are not possible, the measurement results are supplemented by calculations. All calculations are performed by the control unit 50 or an externally connected PC.

FIG. 8 is a flowchart of the pre-processing, and FIG. 9 is a diagram showing the output of the reading unit corresponding to the background board.
First, in step S202, the image sensor 13e is caused to perform a reading operation with the LED 13d turned off to acquire black reference data BK. In step S204, the image sensor 13e is caused to perform a reading operation of the background board 13c with the LED 13d turned on to acquire background board data GR. The background board data GR includes data for the white background portion 13c1 and data for the black background portion 13c2. Then, in step S206, the image sensor 13e is caused to perform a reading operation of the white reference chart to acquire white reference data WH. The white reference data WH is stored in non-volatile memory.

Next, in step S208, the missing measurement results are calculated. First, since the center of the background board 13c is black, the following calculation is performed to create center data.
The difference between the white reference data WH and the white background portion 13c1 of the background board 13c is calculated. This is done by calculating the L difference using equation (6) and the R difference using equation (7) for the left edge data (white) GR_L of the left edge white background portion 13c1 and the right edge data (white) GR_R of the right edge white background portion 13c1, and then calculating them individually. Here, it is sufficient to find the average value for all pixels in the area.
L difference = (white reference data WH_L) - (left edge data of background board (white) GR_L) ... (6)
R difference = (white reference data WH_R) - (right end data of background board (white) GR_R) ... (7)

Next, based on equations (8) and (9), the measurement results for the missing pixels in the central part of the background board 13c against the white background (background board central left data GR_S_L and background board central right data GR_S_R) are calculated.
Background board center left data GR_S_L = white reference data WH (center) - L difference ... (8)
Background board center right data GR_S_R = white reference data WH (center) - R difference ... (9)
In this way, the background board data GR″ is made up of values arranged from the left end as GR_L, GR_S_L, GR_S_R, and GR_R.

In other words, to obtain white shading data, a bright white or gray white background 13c1 area is required for the background board 13c, but this is insufficient in the black background 13c2 area. Therefore, for the black background 13c2, the difference between the measurement results of the white background 13c1 measured at both ends and the white reference chart is calculated, and the measurement results corresponding to the white background are estimated by subtracting this difference from the measurement results of the white reference chart measured in the black background 13c2 area.

In the case of equations (6) to (9), to improve the accuracy of the estimation, the difference value at the left edge and the difference value at the right edge are separated, and for the black background portion 13c2, the difference value at the left edge is applied to the area close to the left edge, and the difference value at the right edge is applied to the area close to the right edge. Instead of illuminating the entire surface, LED 13d is sometimes placed at the right or left edge, and a light guide plate is used to spread the light over the entire surface. In such cases, even if a light guide plate is used, the illuminance on the side opposite the light source tends to be low. For this reason, it is possible to improve accuracy by weighting the difference value at the left edge and the difference value at the right edge according to the distance from the light source and applying it to the central pixel.

In this way, for the area of the background board 13c that is the low-brightness black background portion 13c2, the difference value between the high-brightness white background portion 13c1 and the white reference chart is used to calculate the background board data GR from the measurement results for the white reference chart in the black background portion 13c2.
In this example, white background portions 13c1, 13c1 are arranged at both ends, and black background portion 13c2 is arranged in the center, but background board data can also be obtained for black background portion 13c2 by using the measurement results of white background portion 13c1, so various variations are possible with regard to the arrangement of white background portion 13c1 and black background portion 13c2, which will be described later.
However, if the white background portion 13c1 serving as the white reference is provided on both ends, the amount of acquired white reference data can be increased compared to when it is provided on only one side, leading to improved shading correction.

Thereafter, in step S210, the background board RAW data is calculated based on the following equation (10), and in step S212, the white reference RAW data is calculated based on the following equation (11):
Background board RAW data = GR" - BK ... (10)
White reference RAW data = WH - BK ... (11)
It is stored in non-volatile memory and shipped from the factory.

FIG. 10 is a flowchart showing the process of calculating the white shading data.
Immediately before scanning, white shading data is calculated as follows: In step S222, the image sensor 13e is caused to perform a reading operation with the LED 13d turned off to acquire black reference data BK_S, and in step S224, the image sensor is caused to perform a reading operation of the background board 13c with the LED 13d turned on to acquire background board data GR_S. The background board data GR_S is calculated in the same manner as in equations (6) to (9). At this time, white reference data WH stored in non-volatile memory is used.
Thereafter, in step S226, the previous background board RAW data is calculated based on equation (12).

Previous background board RAW data = GR_S - BK_S ... (12)
Furthermore, in step S228, calculation is performed based on equation (13).
White shading data =
(White reference RAW data) / (Background board RAW data) x Previous background board RAW data ... (13)
As described above, according to the above-described example, shading correction and document cutout can be performed appropriately without providing a mechanism for switching between white and black on the background plate, and the number of parts can be reduced.

As such, the image reading device described above is equipped with a control unit 15 that controls image reading. This control unit 15 reads the white background portion 13c1 and the black background portion 13c2 using the image sensor 13e, which serves as the reading unit, when no original G is present, and generates correction data for correcting the read data of the original G based on the read data of the white background portion 13c1 and the black background portion 13c2.

As described above, the black background portion 13c2 placed in the center is wider than the width of the assumed main document G, and if the assumed main document G is placed exactly in the center in the width direction of the transport path and transported straight, parts of the black background portion 13c2 will appear on both sides of the document G in the width direction, allowing the image to be accurately cut out.
However, the document G does not necessarily pass through the center of the transport path straight and without tilting as expected. The document extraction process using the black background portion 13c2 disposed on the background plate 13c will now be described.

When both ends of the background board are white and the center is black, the background board 13c has white background portions 13c1, 13c1 arranged on the outer sides in the width direction, with a black background portion 13c2 arranged between them. The width of the black background portion 13c2 is wider than the width of the expected main document G. When the size is like this, the pixels included in the white areas at both ends of the background board 13c in the width direction are excluded from the processing target, and the document is cut out from the black area.

FIG. 11 is a diagram showing a first pattern of document cutting processing.
1) If the document G is sized within the black background portion 13c2, the four vertices would normally be included in the black region in the scanned image data.
1: Extract the parts that form vertices within the black area.
2: If the number of extracted vertices is four, the line formed by connecting the four vertices is identified as the black boundary, and its slope is calculated.
3: The image is cut out using the black border of the black background and tilt correction is performed.
As shown on the left side of FIG. 11, the entire document G is included within the black area, and as shown on the right side of the same figure, an image formed by connecting four vertices is clipped after being tilt-corrected.

2) When the document G overlaps the white area of the background plate 13c FIG. 12 is a diagram showing a second pattern of the document cutting process.
(Part 1)
1: Extract the parts that form vertices within the black area.
2: If the number of vertices in the black area is three, identify the black boundary that forms the top side and determine the slope.

After this, two extraction methods are possible.
3-1: The coordinates of the remaining vertex in the white area are found based on the coordinates of the three vertices. The image contained in the four vertices is tilt-corrected and cut out. The process of tilt-correcting and cutting out the image contained in the four vertices is the same as the method shown in Figure 11.
3-2: Cut out the part of the image that extends into the white area, correct the tilt, and cut out. By cutting the image, it becomes a distorted pentagon, but it does not look strange because the background is white.
As shown on the left side of Figure 12, three vertices are included in the black area, and when four vertices are assumed as shown in the center of the figure, the image of the white area is assumed to be absent, and only the part included in the black area is corrected for tilt and cut out as shown on the right side of the figure.

In this way, if three of the four vertices of document G are within the black area, the control unit 15 will not use the scanned data of document G that extends beyond the black area. By doing this, the accuracy of document extraction can be improved and the amount of data can be reduced.

(Part 2)
If there are no vertices in the black region or if there are one or two vertices in the black region, it is difficult to accurately determine the edge of the image because there are not three vertices in the black region.
13 shows a third pattern of document extraction processing. As shown in the upper part of the figure, both the left and right sides of the image overlap the white area. Since the four corners are in the white area, the exact position of the image cannot be determined, but it is definitely within the paper transport path.

1: Assume that the document is shifted to the leftmost side of the transport path. This is shown on the left side of the lower part of FIG.
2: If we assume that the document G is shifted to the left side of the transport path due to the inclination of the black boundary that forms the upper edge near the top end of the black area, the intersection of the line extending the upper edge black boundary to the left and the left end of the transport path can be considered as one vertex.
3: By taking the slope of the black boundary on the top edge and the assumed intersection point as one vertex, and finding the black boundary that forms the bottom edge that appears near the bottom end of the black area, the length of the image can be calculated.
4: After determining the length of the image, determine the side that intersects the top side at a 90-degree angle from the vertex of the intersection previously determined, and identify the vertex to the left of the bottom side from the determined length of the image. Then, determine the distance L1 from this vertex to the left edge of the white area.
5: Regarding the right end, if we assume that the document G is shifted to the right side of the transport path due to the inclination of the black boundary that forms the lower edge that appears near the bottom end of the black area, the intersection of the line extending the black boundary of the lower edge to the right and the right end of the transport path can be considered as one vertex.
6: The intersection point on the right side of the bottom side is set as the vertex, and a side is assumed that intersects the bottom side at a 90-degree angle from the intersection point. The vertex on the right side of the top side is identified from the calculated image length. Then, the distance L2 from this vertex to the right edge of the white area is determined.
7: The left side of the bottom row of Figure 13 shows an image surrounded by the four identified vertices, and as shown in the center of the figure, on the left side, a section with a width of length L1 is cut from the left end, and on the right side, a section with a width of length L2 is cut from the right end.
8: Finally, the image with both sides cut off is corrected for the tilt calculated in advance and cut out. At this time, the cut-out images on both sides may contain white background, but since it is white, it does not look strange.

As described above, the image reading device is equipped with a control unit 15 that controls the reading of the image, and this control unit 15 generates correction data for correcting the reading data of the original G as described above, and further, after reading the image of the original G, cuts out the image data of the original based on the reading data of the original G in the black area.
Therefore, the accuracy of document extraction can be improved with a simple background plate structure in which only a part of the same fixed surface is made into the white background portion 13c1 and the rest is made into the black background portion 13c2.

Furthermore, if two or fewer of the four vertices of document G are within the black area, the control unit 15 identifies the vertex at the leading or trailing edge of document G based on the intersection of an extension of the edge (black boundary) of the leading or trailing edge of the document passing through the black area and the edge (left or right edge) of the reading area of the image sensor 13e of the reading unit, which is a white area, and does not use the read data for the area outside the reading area based on the distances L1 and L2 obtained using this vertex as a reference.

As described above, the background plate 13c does not necessarily have to have the white background portions 13c1, 13c1 located at both ends and the black background portion 13c2 located in the center.
FIG. 14 is a schematic diagram showing a background plate and a reading unit according to another example.
As shown in the figure, compared to the background plate 13c described above, the white background portion 13c1 located on the right has been eliminated, and the black background portion 13c2 has been extended to this portion. When the maximum effective pixel width of the image sensor 13e is used as the reference, the reading element at the left end of the image sensor 13e faces the white background portion 13c1, and the reading element at the right end faces the black background portion 13c2.
One end of the background plate 13c is white (either only the left side or only the right side), and the remaining part of the background plate 13c excluding the one end is black.

If the scanned image data contains four vertices within a black area, the straight line connecting the four vertices is identified as the black boundary, and its inclination is corrected before extraction.

When the document G overlaps the background board in the white area When there are three vertices in the black area FIG. 15 is a diagram showing a fourth pattern of the document cutting process.
In this case, document extraction may be performed when the document G overlaps the background board in the white region and three vertices are included in the black region, as shown in (No. 1).
FIG. 16 is a diagram showing a fifth pattern of the document cutting process.
In this case, the document can be extracted when the document G overlaps the background board in the white area, and the above-mentioned "when there are no vertices in the black area or when there are one or two vertices in the black area (case 2)" is applied.

FIG. 17 is a schematic diagram showing a background plate and a reading unit according to another example.
As shown in the figure, compared to the background plate 13c described above, the boundary between the white background portions 13c1 located on the left and right and the black background portion 13c2 located in the center has a gradation that is just barely difficult to detect edges with image processing. Because the gradation itself cannot be extracted by edge detection, the edges of the document can be extracted by edge detection.
In this way, a gradation background portion in which the brightness gradually changes is provided between the white background portion 13c1 and the black background portion 13c2.

FIG. 18 is a diagram showing a sixth pattern of the document cutting process.
If the black region has four vertices, the image can be cut out as is, and if the black region has three vertices, the method shown in (Part 1) above can be applied.

When the document overlaps the background of the gradation area When there are two vertices in the black area and the gradation area 1: The slope is calculated from the black boundary on the upper side that appears at the top end of the black area.
2: Find the vertices that lie within the gradient region.
3: An image is extracted from two vertices in the black area and two vertices found by edge detection, and the image is extracted after tilt correction. As shown in Figure 18, the two vertices on the left side of document G are in the gradation area, and the two vertices on the right side are in the black area. Therefore, four vertices can be identified.
If the four vertices of the original G completely overlap the white area of the background board, the above-mentioned "if there are no vertices in the black area or if there are one or two vertices in the black area (case 2)" can be applied.

FIG. 19 is a schematic diagram showing a background plate and a reading unit according to another example.
The background board 13c shown in the figure has one end (either only the left side or only the right side) white as shown in Figure 14 and the remaining end black, and the boundary between the white background portion 13c1 and the arranged black background portion 13c2 is a gradation that is just barely enough to make edge detection possible using image processing.

If there are vertices in the black area and the gradient area: 1: Find two vertices in the black area.
2: Find the vertices that exist within the gradation region using edge detection.
3: The slope is calculated from the black boundary on the upper side that appears at the upper end of the black area.
4: Correct the tilt and cut out the image enclosed by the four vertices found.

When there is a vertex in the white area, the previously described "when there is no vertex in the black area or when there are one or two vertices in the black area (case 2)" can be applied.

FIG. 20 is a schematic diagram showing a background plate and a reading unit according to another example.
As shown in the figure, the background plate 13c can also be made into a black and white stripe.
In this way, a plurality of white background portions 13c1 and black background portions 13c2 are provided, alternately arranged.

FIG. 21 is a diagram showing a sixth pattern of the document cutting process.
If there are no vertices in the black area or if there are one or two vertices in the black area, you can perform the same processing as in "If there are no vertices in the black area or if there are one or two vertices in the black area (case 2)".
When there are three vertices in the black region, the document G can be cut out by correcting the tilt using two methods, as shown in (1) above.

FIG. 22 is a schematic diagram showing a background plate and a reading unit according to another example.
As shown in the figure, the background plate 13c is made of black and white stripes, and the boundary between the white background portion 13c1 and the arranged black background portion 13c2 is made to have a gradation that barely makes it possible to detect the edge by image processing.

If the vertex is in a black area or a gradation area: 1. The slope is determined from the black boundary on the upper side that appears at the upper end of the black area.
2: Find the vertices that exist within the gradation region using edge detection.
3: Find the vertex in the black region.
4: Correct the tilt and cut out the image enclosed by the four vertices found.
If there is a vertex in the white area, the same processing as in "if there is no vertex in the black area or if there are one or two vertices in the black area (case 2)" may be performed.
In this way, by preparing a background plate with a black background in the center and white backgrounds on both ends aligned in the main scanning direction, it is not necessary to switch between the white and black areas of the background plate, and shading correction and image extraction are possible with an inexpensive configuration. In particular, the black center of the background plate is useful for detecting defects in the original document.

(3) Specific examples of image reading devices to which this technology can be applied:
1 to 3 can be applied to the image reading device 10. Therefore, the description already given will be omitted. Here, the details of the image reading device 10 will be described with reference to FIG. 23 onwards.
FIG. 23 schematically illustrates the configuration of the reading unit 13, along with the configurations of the pair of transport rollers 12c and the pair of discharge rollers 12d. The pair of transport rollers 12c includes a transport drive roller 12c1 and a transport driven roller 12c2. The pair of discharge rollers 12d includes a discharge drive roller 12d1 and a discharge driven roller 12d2. FIG. 24 schematically illustrates the configuration of the illumination unit 20 included in the reading unit 13. FIG. 25 schematically illustrates the positional relationship between multiple reading elements 30 aligned along the alignment direction D2 and a background plate 13c in which the white background portion 13c1 is positioned farther from the LED 13d than the black background portion 13c2 in the alignment direction D2. Note that the lower reading unit 13a and the upper reading unit 13b have point-symmetrical structures in FIG. 23 and share the same basic configuration. Therefore, the following description focuses on the configuration of the upper reading unit 13b.

The reading unit 13a includes a transmission plate 21, a background plate 13c, an illumination unit 20, a lens 24, and an image sensor 13e. As shown in Figures 24 and 25, the illumination unit 20 includes an LED 13d arranged on one side of the multiple reading elements 30 in the alignment direction D2, and a light guide 23 that guides light LT1 from the LED 13d to the reading range AR1 of the image sensor 13e. The reading unit 13a is elongated in the alignment direction D2, and the transmission plate 21, background plate 13c, light guide 23, lens 24, and image sensor 13e are also elongated in the alignment direction D2. The reading unit 13a, including the background plate 13c, illumination unit 20, and image sensor 13e, is positioned so that its position in the transport direction D1 does not change. The irradiation unit 20, which irradiates the reading range AR1 with light LT1, irradiates the transported document G with light LT1, and the image sensor 13e reads the image of the transported document G.

The transparent plate 21 can be made of colorless and transparent glass, a resin plate such as a colorless and transparent acrylic plate, etc. The outer surface of the transparent plate 21, i.e., the surface facing the lower reading unit 13a, is a reading surface that reads the document G and forms a transport path CO1 for the document G.
The background plate 13c faces the multiple reading elements 30 of the reading unit 13a at a position where it is irradiated with light LT1 from the irradiation unit 20 of the opposing lower reading unit 13a, and is read by the multiple reading elements 30 of the reading unit 13a for shading correction. The surface of the background plate 13c facing the lower reading unit 13a is the reading reference surface 22a. As shown in FIG. 25 , the background plate 13c includes a white background portion 13c1 for shading correction and a black background portion 13c2 for document cutout that is lower in brightness than the white background portion 13c1 on the reading reference surface 22a. The black background portion 13c2 and the white background portion 13c1 are aligned in the alignment direction D2, and the white background portion 13c1 is positioned farther from the LED 13d than the black background portion 13c2 in the alignment direction D2. The length of the black background portion 13c2 in the alignment direction D2 is longer than the length of the white background portion 13c1 in the alignment direction D2. The area of the black background portion 13c2 is larger than the area of the white background portion 13c1.

24, the LED 13d faces the incident end 23a, which is one end of the light guide 23 whose longitudinal direction faces the arrangement direction D2, and emits light LT1 toward the incident end 23a. As shown in FIG. 25, the LED 13d may include three types of LEDs: an LED that emits red (R) light, an LED that emits green (G) light, and an LED that emits blue (B) light.
The light guide 23 is a long light-guiding member for irradiating light LT1 toward the document G or background plate 13c. As shown in FIG. 24 , the light guide 23 guides light LT1 emitted from the LED 13d in the alignment direction D2 toward the opposing lower reading unit 13a. Light LT1 from the LED 13d enters the light guide 23's incident end 23a. Light LT1 exits from an exit surface 23c of the light guide 23 facing the lower reading unit 13a. The light guide 23's reflective surface 23b, located opposite the exit surface 23c, is roughened in the alignment direction D2 by increasing in roughness with increasing distance from the LED 13d, so that the light LT1 emitted from the exit surface 23c is substantially uniform. Light LT1 entering the light guide 23 from the incident end 23a is diffusely reflected by the reflective surface 23b and exits from the exit surface 23c. Light emitted from the light guide 23 toward the lower reading unit 13a is reflected by the document G or the background plate 13c of the reading unit 13a, reaches the image sensor 13e via the lens 24, and is photometrically measured by the image sensor 13e.

The image sensor 13e is equipped with a large number of reading elements 30, which are photoelectric conversion elements. The arrangement density of the reading elements 30 is not particularly limited, but can be, for example, equivalent to 300 to 600 dpi. The reading elements 30 can be, for example, photodiodes. One reading element 30 corresponds to one pixel, and the image sensor 13e outputs the reading results obtained by photometry performed by each reading element 30 to the control unit 15. The image sensor 13e can be a Contact Image Sensor type (abbreviated as CIS type) or a Charge Coupled Device type (abbreviated as CCD), or a solid-state imaging element such as a CMOS image sensor or a line sensor or area sensor composed of a CCD. Here, CMOS is an abbreviation for Complementary Metal-Oxide Semiconductor.

The image sensor 13e shown in Figure 25 includes multiple sensor chips 13f aligned along the alignment direction D2. Each sensor chip 13f includes multiple reading elements 30 aligned along the alignment direction D2. As a result, the image sensor 13e has a large number of reading elements 30 across a wide reading range AR1 in the alignment direction D2 that exceeds the maximum effective pixel width W1, which is the standard size defined by ISO 216 plus a margin. Here, ISO is an abbreviation for International Organization for Standardization, and the standard size is, for example, A4 or A3 size, with the margin being a size that allows for skew in the transported document G. The reading range AR1 is the maximum range over which the multiple reading elements 30 read the background board 13c in the alignment direction D2. The maximum effective pixel width W1 is the maximum range over which the image sensor 13e reads the image of the document G in the alignment direction D2, and is slightly narrower than the reading range AR1, a predetermined range that is the standard size plus a margin.

In the alignment direction D2, the range of the multiple reading elements 30 facing the background plate 13c is narrower than the combined ranges AR2 and AR3 of the black background portion 13c2 and the white background portion 13c1. Here, among the multiple reading elements 30 included in the image sensor 13e, the multiple reading elements located within the maximum effective pixel width W1 are referred to as the multiple first elements 31, and the multiple reading elements located outside the maximum effective pixel width W1 are referred to as the multiple second elements 32. In the alignment direction D2, the black background portion 13c2 includes the maximum effective pixel width W1. Therefore, the multiple first elements 31 are within the range AR2 of the black background portion 13c2 in the alignment direction D2. In the alignment direction D2, the range of the multiple first elements 31 facing the black background portion 13c2 is narrower than the range AR2 of the black background portion 13c2. Furthermore, in the alignment direction D2, the white background portion 13c1 is outside the maximum effective pixel width W1, which is the standard size defined in ISO 216 plus a margin. In the alignment direction D2, the range of the multiple second elements 32 facing the background plate 13c is a range that includes at least a portion of the white background portion 13c1. Figure 25 shows that the range of the multiple second elements 32 in the alignment direction D2 is a range that includes a portion of the black background portion 13c2 and a portion of the white background portion 13c1. In other words, the end 13c4 of the black background portion 13c2 on the white background side in the alignment direction D2 is slightly outside the maximum effective pixel width W1. This is to ensure that the multiple first elements 31 present within the maximum effective pixel width W1 face the black background portion 13c2, taking into account the installation tolerance of the reading unit 13. Of course, in the alignment direction D2, all second elements 32 present outside the maximum effective pixel width W1 may face the white background portion 13c1.

As illustrated in FIG. 31 , the output characteristics of the multiple reading elements 30 are not uniform. FIG. 31 schematically illustrates output characteristics 901 and 902 with respect to the position X of the reading element 30 in the alignment direction D2. The position X of each reading element 30 can also be considered the position of the corresponding pixel. The vertical axis in FIG. 31 indicates the read value L measured when the reading unit 13 is provided with a background plate 13c in which all reading elements 30 included in the image sensor 13e face the white background portion 13c1. The read value L is expressed, for example, in luminance units of LSB. Output characteristic 901 indicates the output characteristic when the distance between the background plate 13c and the multiple reading elements 30 is relatively short, and output characteristic 902 indicates the output characteristic when the distance between the background plate 13c and the multiple reading elements 30 is relatively long.
In both output characteristics 901 and 902, the read value L tends to decrease as the distance from the LED 13d increases in portions of the plurality of reading elements 30 farther from the LED 13d, particularly in portions within the reading range AR1 that are outside the maximum effective pixel width W1. However, as the distance between the background board 13c and the plurality of reading elements 30 increases, uniformity deteriorates, and the read value L tends to decrease as the distance from the LED 13d increases in portions of the plurality of reading elements 30 farther from the LED 13d.

In this specific example, therefore, by utilizing multiple second elements 32 located outside the maximum effective pixel width W1, virtual background board read data is obtained assuming that the black background portion 13c2 is the white background portion 13c1, and correction data for shading correction is obtained. Here, if a sensor chip 13f for reading the white background portion 13c1 outside the maximum effective pixel width W1 were added to the image sensor 13e, the image sensor 13e would become larger in the alignment direction D2, which would also lead to increased costs for the image sensor 13e. In this specific example, there is no need to add a sensor chip 13f to the image sensor 13e, so the image reading device 10 does not become larger in the alignment direction D2.

FIG. 26 illustrates a schematic example of the relationship between the position X of the reading element 30 in the alignment direction D2 and the read value L.
The black reference read data BK is read data of the background board 13c not irradiated with light LT1, i.e., read data of the background board 13c when light LT1 is not irradiated from the irradiation unit 20. The background board read data GR is read data of the background board 13c irradiated with light LT1, i.e., read data of the background board 13c when light LT1 is irradiated from the irradiation unit 20. In the background board read data GR, the read values L of the multiple first elements 31 facing the black background portion 13c2 are relatively small, and the read values L of the multiple second elements 32 facing the white background portion 13c1 are relatively large. The white reference read data WH is read data of a white reference chart, such as a white reference board, placed within the maximum effective pixel width W1 and irradiated with light LT1. The white reference chart can be considered an original G having a substantially pure white image. The white reference read data WH may be the read values L of all the read elements 30 present in the read range AR1, but it is sufficient to have the read values L of multiple first elements 31 present in the maximum effective pixel width W1. Note that the read data means a collection of the read values L of each read element 30 according to the position X.

Because the multiple first elements 31 within the maximum effective pixel width W1 face the black background portion 13c2, the background plate read data GR cannot be used directly for shading correction. Therefore, as illustrated in Figures 27 and 28, the background plate read data GR is converted into virtual background plate read data GR', which corresponds to the read data of the background plate 13c illuminated by light LT1, assuming that the black background portion 13c2 is the white background portion 13c1.

Fig. 27 shows a schematic diagram of how the background board read data GR is converted into virtual background board read data GR' corresponding to the read data of the white background portion 13c1. Fig. 28 shows a schematic diagram of how the background board read data GR is converted into virtual background board read data GR' based on the difference between the black background read value Lb and the virtual white background read value L0.
As shown in Fig. 31, the output characteristics 901, 902 of the multiple reading elements 30 tend to have a smaller read value L the farther away from the LED 13d in the portion of the reading range AR1 outside the maximum effective pixel width W1. Therefore, first, based on multiple white background read values Lw1, Lw2 at multiple reading positions in the white background portion 13c1 in the arrangement direction D2, a virtual white background read value L0 is calculated, which is assumed to be the read value at a specific position on the boundary B1 of the maximum effective pixel width W1 in the arrangement direction D2. In the example shown in Fig. 28, pixel n0 is located at the specific position, and pixels n1 and n2 are located at multiple reading positions. The specific position on the boundary B1 of the maximum effective pixel width W1 is not limited to the position of the most extreme first element among the plurality of first elements 31, but may be the position of a first element in the black background portion 13c2 that is closer to the end 13c4 on the white background portion side than the end 13c3 on the light source side, among both ends in the arrangement direction D2. The number of reading positions is not limited to two, and may be three or more.

In this specific example, a virtual white background read value L0 at a specific position (pixel n0) is obtained by performing an approximation calculation based on the relationship between multiple white background read values Lw1 and Lw2 for multiple reading positions (pixels n1 and n2). The approximation of the relationship between multiple white background read values Lw1 and Lw2 for multiple reading positions may be a polynomial approximation of second or higher order, or may be a linear approximation. Figure 28 shows an example of obtaining a virtual white background read value L0 using linear approximation. Here, in the alignment direction D2, the position of pixel n0 is represented by n0, the position of pixel n1 is represented by n1, and the position of pixel n2 is represented by n2, with 0 < n0 < n1 < n2. The virtual white background read value L0 can be calculated using the following formula.
L0=Lw2+(Lw1-Lw2)/(n2-n1)×(n2-n0)...(14)

Here, the read value L of pixel n0 at a specific position on boundary B1 is defined as black background read value Lb. The difference L0-Lb between the black background read value Lb at pixel n0 and the virtual white background read value L0 corresponds to the difference between the read data of black background portion 13c2 and the read data of white background portion 13c1 when light LT1 is irradiated onto background board 13c. Therefore, the virtual background board read data GR' is obtained by adding the aforementioned difference L0-Lb as an offset to the portion of background board read data GR that corresponds to maximum effective pixel width W1.
GR'=GR+(L0-Lb)...(15)
As described above, the virtual white background read value L0 at a specific position can be obtained by performing approximate calculations based on the relationship between the plurality of white background read values Lw1 and Lw2 for the plurality of read positions in the alignment direction D2.

(4) Specific examples of processing by an image reading device:
29 shows a schematic example of the pre-processing in which background plate raw data GR'-BK and white reference raw data WH-BK are calculated and stored in non-volatile memory 17. This pre-processing is performed by control unit 15 before shipping image reading device 10. Control unit 15 starts the pre-processing when a predetermined operation to start the pre-processing is received, for example, via touch panel 16. Hereinafter, step symbols may be shown in parentheses.

When the pre-processing begins, the control unit 15 causes the image sensor 13e to read the background board 13c with the LED 13d turned off, and acquires the black reference reading data BK, which is the reading data of the background board 13c that is not illuminated by the light LT1, from the image sensor 13e (step S302).
Next, the control unit 15 turns on the LED 13d and then causes the image sensor 13e to read the background board 13c, and acquires background board read data GR, which is read data of the background board 13c illuminated by the light LT1, from the image sensor 13e (step S304).

The operator then places the white reference chart within the range where the image sensor 13e reads the document image. In this state, the control unit 15 causes the image sensor 13e to read the white reference chart while the LED 13d is lit, and acquires white reference read data WH, which is the read data of the white reference chart illuminated by light LT1, from the image sensor 13e (step S306).

Next, the control unit 15 obtains, from the background board read data GR obtained in step S304, the black background read value Lb of pixel n0 located in boundary B1 of the maximum effective pixel width W1, and the white background read values Lw1 and Lw2 of pixels n1 and n2 corresponding to multiple read positions in the white background portion 13c1 (step S308). The black background read value Lb is the read value L at a specific position in the black background portion 13c2 that is closer to the end 13c4 on the white background side than to the end 13c3 on the light source side. The white background read values Lw1 and Lw2 are the read values L at multiple read positions in the white background portion 13c1 in the alignment direction D2.

Next, the control unit 15 calculates a virtual white background read value L0 based on the multiple white background read values Lw1 and Lw2, which is assumed to be the read value L at a specific position on the background board 13c, assuming that the specific position on the boundary B1 of the maximum effective pixel width W1 is the white background portion 13c1 (step S310). The processing of step S310 is performed because the portion of the background board 13c corresponding to the maximum effective pixel width W1 is the black background portion 13c2. As shown in FIG. 28, the control unit 15 obtains the virtual white background read value L0 at the specific position by performing an approximate calculation based on the relationship between the multiple white background read values Lw1 and Lw2 for the multiple read positions in the alignment direction D2. For example, the control unit 15 can calculate the virtual white background read value L0 according to the above-mentioned equation (14).

Next, the control unit 15 creates virtual background board read data GR' corresponding to the read data of the white background portion 13c1 by adding the difference L0-Lb between the virtual white background read value L0 and the black background read value Lb as an offset to the background board read data GR for the maximum effective pixel width W1 (step S312). In this way, the control unit 15 converts the background board read data GR into virtual background board read data GR' based on the difference L0-Lb between the black background read value Lb and the virtual white background read value L0.

Next, the control unit 15 calculates background board RAW data GR'-BK, which represents the difference between the virtual background board read data GR' and the black reference read data BK, and stores the background board RAW data GR'-BK in non-volatile memory 17 (step S314). As described above, the virtual background board read data GR' corresponds to the read data of the background board 13c illuminated with light LT1 when assuming that the black background portion 13c2 is the white background portion 13c1, and the black reference read data BK is the read data of the background board 13c not illuminated with light LT1. The control unit 15 may also store the virtual background board read data GR' and the black reference read data BK separately as background board RAW data in non-volatile memory 17. The background board RAW data GR'-BK only needs to contain data of the maximum effective pixel width W1.

Finally, the control unit 15 calculates white reference RAW data WH-BK, which represents the difference between the white reference read data WH and the black reference read data BK, and stores the white reference RAW data WH-BK in the non-volatile memory 17 (step S316). As described above, the white reference read data WH is read data from the white reference chart. Note that the control unit 15 may store the white reference read data WH and the black reference read data BK separately as white reference RAW data in the non-volatile memory 17. The white reference RAW data WH-BK only needs to contain data for the maximum effective pixel width W1.

Assuming that the image of the document G is read without shading correction, the control unit 15 of this specific example converts the read value L into an output value so that when the read value L is the value of the black reference read data BK, the output value is set to the minimum value 0, and when the read value L is the value of the white reference read data WH, the output value is set to the maximum value 255, as described with reference to Fig. 6. If the output value is Dout, the output value Dout can be calculated by the following formula.
Dout=(L-BK)/(WH-BK)×255...(16)
In practice, in order to compensate for uneven density between pixels due to uneven illumination, etc., the control unit 15 determines the white shading data SD at least either before reading the document G or when power is turned on to the image reading device 10. The white shading data SD is an example of correction data for shading correction.

Figure 30 shows a schematic example of the white shading correction process. The control unit 15 starts the white shading correction process before reading the original G and when the image reading device 10 is powered on. The control unit 15 may not perform the white shading correction process when the image reading device 10 is powered on, but may perform the white shading correction process before reading the original G. The control unit 15 may also not perform the white shading correction process before reading the original G, but may perform the white shading correction process when the image reading device 10 is powered on. Here, step S402 corresponds to the first step, step S404 corresponds to the second step, step S406 corresponds to the third step, step S408 corresponds to the fourth step, step S410 corresponds to the fifth step, and steps S412 to S414 correspond to the sixth step.

When the white shading correction process starts, the control unit 15 causes the image sensor 13e to read the background board 13c with the LED 13d turned off, and acquires the black reference reading data BK_S, which is the reading data of the background board 13c that is not illuminated by the light LT1, from the image sensor 13e (step S402).
Next, the control unit 15 turns on the LED 13d and then causes the image sensor 13e to read the background board 13c, and acquires background board read data GR_S, which is read data of the background board 13c illuminated by the light LT1, from the image sensor 13e (step S404).

Next, the control unit 15 obtains, from the background board read data GR_S obtained in step S404, the black background read value Lb of pixel n0 located in boundary B1 of the maximum effective pixel width W1, and the white background read values Lw1 and Lw2 of pixels n1 and n2 corresponding to multiple read positions in the white background portion 13c1 (step S406). As described above, the black background read value Lb is the read value L at a specific position in the black background portion 13c2 that is closer to the end 13c4 on the white background side than to the end 13c3 on the light source side. The white background read values Lw1 and Lw2 are the read values L at multiple read positions in the white background portion 13c1 in the alignment direction D2.

Next, the control unit 15 calculates a virtual white background read value L0 based on the multiple white background read values Lw1 and Lw2, which is assumed to be the read value L at a specific position on the background board 13c at the boundary B1 of the maximum effective pixel width W1, assuming that the specific position is the white background portion 13c1 (step S408). As shown in FIG. 28, the control unit 15 obtains the virtual white background read value L0 at the specific position by performing an approximate calculation based on the relationship between the multiple white background read values Lw1 and Lw2 for the multiple read positions in the alignment direction D2. For example, the control unit 15 can calculate the virtual white background read value L0 according to the above-mentioned equation (14).

Next, the control unit 15 creates virtual background board read data GR_S' equivalent to the read data of the white background portion 13c1 by adding the difference L0-Lb between the virtual white background read value L0 and the black background read value Lb as an offset to the background board read data GR_S for the maximum effective pixel width W1 (step S410). In this way, the control unit 15 converts the background board read data GR_S into virtual background board read data GR_S' based on the difference L0-Lb between the black background read value Lb and the virtual white background read value L0.

Next, the control unit 15 calculates the previous background board RAW data GR_S'-BK_S, which represents the difference between the virtual background board read data GR_S' and the black reference read data BK_S (step S412). As described above, the virtual background board read data GR_S' corresponds to the read data of the background board 13c illuminated with light LT1 when it is assumed that the black background portion 13c2 is the white background portion 13c1, and the black reference read data BK_S is the read data of the background board 13c not illuminated with light LT1. The previous background board RAW data GR_S'-BK_S only needs to be data of the maximum effective pixel width W1.

Finally, the control unit 15 calculates white shading data SD based on the previous background board RAW data GR_S'-BK_S, as well as the background board RAW data GR'-BK and white reference RAW data WH-BK stored in the non-volatile memory 17 (step S414). The white shading data SD only needs to be data for the maximum effective pixel width W1. The white shading data SD can be calculated, for example, using the following formula:
SD={(WH-BK)/(GR'-BK)}×(GR_S'-BK_S)...(17)

Equation (17) can also be said to multiply the white reference raw data WH-BK by the ratio {(GR_S'-BK_S)/(GR'-BK)} of the previous background plate raw data GR_S'-BK_S to the background plate raw data GR'-BK. Therefore, the white shading data SD can be said to be the read data of the white reference chart that is assumed when a white reference chart, which may not be available to the user, is placed within the range where the image of the document G is read. The control unit 15 converts the read value L to an output value, for example, by setting the output value to the minimum value of 0 when the read value L is the value of the black reference read data BK_S, and setting the output value to the maximum value of 255 when the read value L is the value of the white shading data SD. Let Dout' be the output value after shading correction. The output value Dout' can be calculated using the following equation:
Dout'=(L-BK_S)/SD×255...(18)
As described above, the control unit 15 can perform shading correction using the white shading data SD.

As explained above, virtual background board read data GR_S' is obtained from background board read data GR_S, which is read data of background board 13c including black background portion 13c2 and white background portion 13c1, assuming that black background portion 13c2 is white background portion 13c1, and white shading data SD for shading correction is obtained. The control unit 15 can perform shading correction using the white shading data SD to cut out the image of document G. Therefore, this specific example eliminates the need for a mechanism to switch between a white background for shading correction and a black background for document cutout. As a result, this leads to a reduction in the number of parts in the image reading device, making it possible to perform shading correction and image cutout with an inexpensive configuration.

(5) Modification:
The present invention can be modified in various ways.
For example, the image reading device is not limited to a double-sided reading device that reads images on both sides of the document G, but may be a single-sided reading device that reads images on only one side of the document G.
The above-described processes can be changed as appropriate, such as by changing the order of the processes. For example, in the pre-processing shown in Fig. 29, the process of S302 can be performed after any of the processes of S304, S306, S308, S310, and S312, as long as it is performed before the process of S314. In the white shading correction process shown in Fig. 30, the process of S402 can be performed after any of the processes of S404, S406, S408, and S410, as long as it is performed before the process of S412.

The arrangement of the white background portion 13c1 and the black background portion 13c2 on the background board 13c is not limited to the arrangement shown in Figures 25 to 28. For example, the background board 13c may include a white background portion 13c1 that is positioned farther from the LED 13d in the alignment direction D2 than the black background portion 13c2, as well as a white background portion 13c1 that is positioned closer to the LED 13d in the alignment direction D2 than the black background portion 13c2. Furthermore, the background boards 13c shown in Figures 4, 17, 20, and 22 also include a white background portion 13c1 that is positioned farther from the LED 13d in the alignment direction D2 than the black background portion 13c2, and therefore are applicable to the present technology.

Furthermore, as illustrated in FIG. 32 , this technology can also be applied when there is a bias in the outputs of multiple reading elements 30 within a single sensor chip 13f. FIG. 32 schematically illustrates an example of the output characteristics of the sensor chip 13f. The position X of each reading element 30 can also be considered the position of the corresponding pixel. The vertical axis illustrated in FIG. 32 represents the read value L measured when the reading unit 13 is provided with a background plate 13c in which all reading elements 30 included in the sensor chip 13f face the white background portion 13c1. Output characteristics 911 represent the output characteristics when the environmental temperature of the sensor chip 13f is relatively low, and output characteristics 912 represent the output characteristics when the environmental temperature of the sensor chip 13f is relatively high.
In both output characteristics 911 and 912, there is a peak PK at the center in the alignment direction D2 where the read value L is the largest, and the read value L tends to decrease as the distance from the peak PK increases. However, as the environmental temperature of the sensor chip 13f increases, uniformity deteriorates, and the read value L tends to decrease as the distance from the peak PK increases.

Therefore, when the reading element 30 at the position of peak PK faces the black background portion 13c2, and among the multiple reading elements 30 included in the sensor chip 13f, there are multiple second elements 32 that face the white background portion 13c1, the control unit 15 can perform the processing shown in Figures 29 and 30. As a result, the control unit 15 can perform shading correction using the white shading data SD to cut out the image of the document G.

(6) Conclusion:
As explained above, according to the present invention, it is possible to provide a technology that eliminates the need for a mechanism for switching between a white background for shading correction and a black background for document extraction, etc. Of course, even if the technology consists only of the constituent elements of the independent claims, the basic functions and effects described above can be obtained.
Furthermore, it is possible to implement configurations in which the components disclosed in the above examples are substituted with each other or the combination is changed, or configurations in which the components disclosed in the publicly known techniques and the above examples are substituted with each other or the combination is changed, etc. The present invention also includes these configurations.

10...image reading device, 11...placement section, 12...transport section, 12a...feed roller, 12b...separation roller, 12c...transport roller, 12d...discharge roller, 13, 13a, 13b...reading unit, 13c...background plate, 13c1...white background section, 13c2...black background section, 13c3...end section on the light source side, 13c4...end section on the white background section side, 13d...LED (example of light source), 13e...image sensor (example of reading section), 13f...sensor chip, 14...detection section, 14a...placement section detection sensor, 14b...document detection sensor, 14c...double feed detection sensor sensor, 15...controller, 16...touch panel, 17...non-volatile memory (example of storage unit), 20...irradiation unit, 23...light guide, 30...reading element, 31...first element, 32...second element, AR1...reading range, AR2...range of black background portion, AR3...range of white background portion, B1...boundary portion, CO1...conveyance path, D1...conveyance direction, D2...alignment direction, G...original, L0...virtual white background read value, Lb...black background read value, Lw1, Lw2...white background read value, LT1...light, SD...white shading data (example of correction data), W1...maximum effective pixel width.

Claims (7)

a reading unit including a plurality of reading elements arranged along an arrangement direction intersecting a document conveyance direction, and configured to read an image of the document being conveyed;
an irradiation unit that irradiates a reading range by the reading unit with light;
a background plate facing the plurality of reading elements at a position where the light is irradiated;
a storage unit for storing data;
a control unit that controls the reading of the image,
the reading unit, the irradiation unit, and the background plate are arranged so that their positions in the conveying direction do not change;
the background plate includes a white background portion for shading correction and a black background portion for document cutout, the black background portion having a lower brightness than the white background portion;
the illumination unit includes a light source disposed on one side of the plurality of reading elements in the arrangement direction, and a light guide that guides the light from the light source to the reading range,
the white background portion is located farther from the light source than the black background portion in the arrangement direction,
The storage unit
background board reference difference data representing the difference between the read data of the background board irradiated with the light and the read data of the background board not irradiated with the light, assuming that the black background portion is the white background portion;
white reference difference data representing a difference between read data of the white reference chart irradiated with the light in a state where the white reference chart is placed in a range where the image is read and read data of the background board not irradiated with the light,
The control unit
acquiring black reference read data, which is read data of the background board to which the light is not irradiated, from the reading unit;
background plate read data, which is read data of the background plate irradiated with the light, is acquired from the reading unit;
From the background board read data, a black background read value at a specific position in the black background portion that is closer to the end of the white background portion side than the end of the light source side among both ends in the arrangement direction, and a plurality of white background read values at a plurality of read positions in the white background portion in the arrangement direction are obtained;
determining a virtual white background read value that is assumed as a read value at the specific position when it is assumed that the specific position is the white background portion based on the plurality of white background read values;
converting the background board read data into virtual background board read data corresponding to the read data of the white background portion based on a difference between the black background read value and the virtual white background read value;
an image reading device that obtains correction data for shading correction based on the difference between the virtual background board read data and the black reference read data, the background board reference difference data, and the white reference difference data; The image reading device of claim 1, wherein the control unit determines the virtual white background reading value at the specific position by performing an approximate calculation based on the relationship between the multiple white background reading values for the multiple reading positions in the alignment direction. The image reading device of claim 1 or claim 2, wherein the control unit performs processing to determine the correction data at least one of before reading the document and when the image reading device is turned on. the plurality of reading elements include a plurality of first elements that are present in a maximum range in which the image is read in the alignment direction, and a plurality of second elements that are present outside the maximum range in the alignment direction,
4. The image reading device according to claim 1, wherein the plurality of first elements are within the range of the black background portion in the arrangement direction. The image reading device of claim 4, wherein the specific position is the position of the boundary of the maximum range in the alignment direction. In the arrangement direction, a range of the plurality of reading elements facing the background plate is narrower than a combined range of the black background portion and the white background portion,
In the arrangement direction, a range of the plurality of first elements facing the black background portion is narrower than a range of the black background portion,
6. The image reading device according to claim 4, wherein a range of the second elements facing the background plate in the arrangement direction includes at least a part of the white background portion. a reading unit including a plurality of reading elements arranged along an arrangement direction intersecting a document conveyance direction, and configured to read an image of the document being conveyed;
an irradiation unit that irradiates a reading range by the reading unit with light;
a background plate facing the plurality of reading elements at a position where the light is irradiated;
a storage unit that stores data,
the reading unit, the irradiation unit, and the background plate are arranged so that their positions in the conveying direction do not change;
the background plate includes a white background portion for shading correction and a black background portion for document cutout, the black background portion having a lower brightness than the white background portion;
the illumination unit includes a light source disposed on one side of the plurality of reading elements in the arrangement direction, and a light guide that guides the light from the light source to the reading range,
the white background portion is located farther from the light source than the black background portion in the arrangement direction,
The storage unit
background board reference difference data representing the difference between the read data of the background board irradiated with the light and the read data of the background board not irradiated with the light, assuming that the black background portion is the white background portion;
a white reference difference data representing a difference between read data of the white reference chart irradiated with light in a state where the white reference chart is placed in a range where the image is read and read data of the background board not irradiated with the light, the white reference difference data including:
a first step of acquiring black reference read data, which is read data of the background plate to which the light is not irradiated, from the reading unit;
a second step of acquiring background plate read data from the reading unit, the background plate being read by the light;
a third step of acquiring, from the background board read data, a black background read value at a specific position in the black background portion that is closer to the end of the white background portion side than the end of the black background portion that is closer to the light source side, and a plurality of white background read values at a plurality of read positions in the white background portion in the arrangement direction;
a fourth step of calculating a virtual white background read value, which is assumed to be a read value at the specific position when it is assumed that the specific position is the white background portion, based on the plurality of white background read values;
a fifth step of converting the background board read data into virtual background board read data corresponding to the read data of the white background portion based on a difference between the black background read value and the virtual white background read value;
a sixth step of determining correction data for shading correction based on the difference between the virtual background board read data and the black reference read data, the background board reference difference data, and the white reference difference data.