JP4808075B2 - Image reading apparatus, program, and recording medium - Google Patents

Description

  The present invention relates to an image reading apparatus that adjusts a reading position in a main scanning direction or a sub-scanning direction by a CIS (contact image sensor) arranged in a predetermined pattern.

  There is a staggered method that realizes A3 and A4 size CIS in a staggered manner at a low cost when reading A0 size CIS. A4 size CIS is low cost because it is produced in large quantities for consumers. On the other hand, the A0 size CIS is less demanded, and the cost of the component part SELFOC lens array and the like is also expensive.

  In the staggered method, each CIS needs to be read in the main scanning direction or the sub-scanning direction as if it were read with a single CIS. Normally, connection adjustment in the main scanning direction is performed by adjusting the reading range of each CIS, and connection adjustment in the sub-scanning direction is performed by adjusting the delay amount of the delay memory.

  However, even with the above adjustment, reading deviation between adjacent CISs remains depending on the accuracy of the mechanical unit, environmental fluctuation, time, and document type. In addition, since the adjustment pitch is determined by the pixel interval and the scanning line interval in the adjustment of the reading range and the delay amount, a shift due to the adjustment pitch remains after the adjustment. Conventionally, the signal processing of the staggered CIS joints (that is, the overlapping reading range) has been performed without taking these deviations into account, so that sufficient correction could not be performed, resulting in image quality degradation.

  The “image reading apparatus” of Patent Document 1 arranges five CISs in a staggered manner, and in order to make the joints of each CIS inconspicuous, the joint part is read redundantly in the adjacent CIS, and the redundantly read pixels A new pixel signal is generated from the signal. When a new pixel signal is generated, a signal of two pixels belonging to different adjacent CISs (confronting each other in the sub-scanning direction) is used, and a new pixel signal for one pixel is generated by an arithmetic expression. The same processing is performed for each two pixels in the overlapping reading range.

However, although it is assumed that each of the two pixels is located on a straight line in the sub-scanning direction (which is perpendicular to the main scanning direction), the image is actually located on an oblique straight line. It was a cause of deterioration. For each of the two pixels, the CIS main scanning line arranged in front of reading is adjusted to adjust the amount of delay so as to eliminate reading deviation in the sub-scanning direction. The deviation remains depending on the document type over time. In the main scanning direction, it is attempted to eliminate reading deviations in the main scanning direction by adjusting the reading range of each CIS. Similarly, deviations remain depending on the accuracy of the mechanical unit, environmental variations, time, and document type. In addition, since the adjustment pitch is determined by the pixel interval and the scanning line interval in the adjustment of the reading range and the delay amount, a shift due to the adjustment pitch remains after the adjustment. In reality, a deviation may remain, and one part (for example, a thin line) may be lost when reading a document image using the above calculation.
JP 2003-46736 A

  In view of the above circumstances, in the present invention, when adjusting the connection of readings by each CIS of the staggered method, the correction of the reading misalignment in the main scanning direction and the sub-scanning direction after the connection adjustment is performed with high accuracy, thereby connecting the CIS joints. Provided is an image reading apparatus or the like that prevents image quality deterioration in a portion (that is, an overlapping reading range). In particular, care should be taken to prevent the missing part of the image and maintain the smoothness of the CIS joint.

To solve the above problems, the image reading apparatus of the present invention, a plurality of image sensors are arranged in a zigzag pattern, and each of the image sensor, the effective read range of the neighboring image sensor in the main scanning direction by a predetermined amount Duplicate in the arrangement image reading apparatus so as to have overlapping effective reading range, and the first thin line of a plurality of parallel in the sub-scanning direction, the angle a plurality of second having with respect to the main scanning direction and the sub scanning direction A contact glass on which a thin line is formed, and a detection range corresponding to an outside of the overlapping effective reading range of an adjacent image sensor among the plurality of image sensors in the main scanning direction is obtained by reading the first thin line . Acquisition means for acquiring a signal interval of the second image sensor, and the detection range of an adjacent image sensor among the plurality of image sensors is the second thin line A setting for setting a delay time when the image sensor on the upstream side in the document conveying direction combines the data obtained by reading the document based on the interval of the second signal obtained by reading Means, delay means for delaying the data read by the upstream image sensor based on the delay time when reading the document with the plurality of image sensors, and the overlapping effective reading range of the respective image sensors . Among them, the signal from which the overlap reading range corresponding to the inside of the main scanning direction of each image sensor is obtained is not the overlap reading range within the overlap reading range based on the interval of the first signal acquired by the acquisition unit. as the amount of movement of the signal read by the side close to the portion is smaller than the image sensor end side in the overlapping reading range, and gradually move And moving means for moving the signal obtained from the overlap reading range in the main scanning direction, and the overlap reading range of each of adjacent image sensors among the plurality of image sensors, and by the delay means characterized by comprising synthesizing means for synthesizing the signal is moved by the moving means and the delay has been performed, the.

According to a second aspect of the present invention, in the image reading apparatus according to the first aspect, the movement in the main scanning direction by the moving unit is obtained by the adjacent image sensor reading the same position in the main scanning direction. The two pixels are moved so as to be in the same position in the main scanning direction.

According to a third aspect of the present invention, in the image reading apparatus according to the first or second aspect, the total of the movement amounts of one of the adjacent image sensors by the moving unit is equal to the total of the movement amount of the other. It is characterized by.

According to a fourth aspect of the present invention, in the image reading apparatus according to any one of the first to third aspects, the synthesizing unit is configured to output each pixel signal based on signals output from the two adjacent image sensors. A determining means for determining the level is provided.

According to a fifth aspect of the present invention, in the image reading apparatus according to any one of the first to third aspects, the portion of the image sensor that is not the overlapping reading range to a predetermined position within the overlapping reading range. Is characterized by using a signal output from the combining means based on the output from each of the image sensors.

According to a sixth aspect of the present invention, in the image reading apparatus according to the fifth aspect, the predetermined position can be changed.

  A seventh aspect of the present invention is an image forming apparatus including the image reading apparatus according to any one of the first to sixth aspects.

The image reading apparatus of the present invention, a plurality of image sensors are arranged in a zigzag pattern, and each of the image sensor, so as to have overlapping effective read range effective reading range of the neighboring image sensor in the main scanning direction by a predetermined amount Duplicate A plurality of first thin lines parallel to the sub-scanning direction and a contact glass on which a plurality of second thin lines having an angle with respect to the main scanning direction and the sub-scanning direction are formed. , Among the plurality of image sensors, the detection ranges corresponding to the outside of the overlapping effective reading ranges of the adjacent image sensors in the main scanning direction acquire the intervals of the first signals obtained by reading the first thin lines , respectively . Among the image sensors, the detection ranges of adjacent image sensors are based on the intervals of the second signals obtained by reading the second thin lines, respectively. In other words, when the image sensor on the upstream side in the document transport direction combines the data obtained by reading the document, a delay time is set, and when reading the document with a plurality of image sensors, the upstream image is based on the delay time. The data read by the sensor is delayed, and the signal from which the overlapping reading range corresponding to the inner side in the main scanning direction of each image sensor is obtained among the overlapping effective reading ranges of the respective image sensors is set to the interval between the acquired first signals. Based on this, the moving amount is gradually changed so that the moving amount of the signal read on the side closer to the portion not in the overlapping reading range in the overlapping reading range becomes smaller than the end side of the image sensor in the overlapping reading range. The signal obtained from the overlapping reading range is moved in the main scanning direction and obtained from each overlapping reading range of the adjacent image sensors among the plurality of image sensors. It is, by combining signals move delay has been performed, it is possible to accurately correct the read offset.

  Hereinafter, the best mode for carrying out the image reading apparatus of the present invention will be described. In the description, the drawings submitted at the same time as this specification will be referred to as appropriate.

  FIG. 1 is a cross-sectional view of the image reading apparatus, and FIG. 2 is a top view of the image reading apparatus, showing an outline of the mechanism. An example using CIS for an image sensor will be described.

  As shown in FIG. 2, three CIS1, CIS2, and CIS3 are arranged in a staggered pattern. In FIG. 1, the document is inserted from the right. When the document insertion sensor 1 detects a document, a transport motor (not shown) rotates a transport roller: front 2 and a transport roller: rear 6 to turn on a light source for document illumination (not shown) in the CIS. . Thereafter, the registration sensor 3 detects the leading edge of the document and sequentially recognizes the position of the document in the conveyance path. The original image (lower surface) is read by CIS2 and then read by CIS1 and CIS3.

  The CIS2 read signal is delayed by the time required for sub-scanning between CIS2 and CIS1, the CIS3 read signal is delayed by the time required for sub-scanning between CIS3 and CIS1, and the CIS1, CIS2, and CIS3 signals are combined, A one-line read signal (inline read signal) is obtained. The document image is sequentially read and discharged through the transport roller 6. The image reading apparatus according to the present embodiment includes a central processing unit (CPU) that performs central processing control of operations in the above-described units, and a ROM (ROM) as a recording medium that records a program that the CPU reads to perform central processing control. Read Only Memory) is installed.

  In FIG. 2, the document is conveyed from top to bottom (when the operation is viewed from the top). The effective reading range of each CIS overlaps with the effective reading range of the adjacent CIS. The effective reading range of each CIS has an effective reading range for obtaining an image signal to be used at each time point. Each effective reading range has a constant reading width (in the main scanning direction) and within the effective reading range (in the main scanning direction). In FIG. 2, left and right direction) can be arbitrarily set independently. This setting is performed by setting the reading start position P1a for CIS1, P2a for CIS2, and P3a for CIS3.

  Each execution reading range has a document reading range, a left position detection range, and a right position detection range. Each document reading range has a left overlapping reading range and a right overlapping reading range. When the effective reading range is moved, the document reading range including the left and right overlapping reading range and the left and right position detection range move in the same amount and in the same direction. In this embodiment, the CIS1 left position detection range, left overlap reading range, CIS3 right position detection range, and right overlap reading range are not used. In addition, the document reading ranges do not necessarily have to be overlapped, but in the present embodiment, description will be made on the case where they overlap.

  A joint using the duplicated signal (for every two sets of joints) is corrected (described later) to obtain an inline read signal. In this reading, it is necessary that the positional relationship between the main scanning direction and the sub-scanning direction of CIS1, CIS2, and CIS3 is accurately adjusted, and even if there is a shift (residual error) after the adjustment, image quality deterioration is not caused. Hereinafter, adjustment of the positional relationship and prevention of image quality deterioration (correction method) will be described.

  The contact glass 5 (or the reading back member 4, hereinafter referred to as the contact glass 5) in FIGS. 1 and 2 is a thin line L1a parallel to the sub-scanning direction at a position corresponding to each CIS left / right position detection range. Fine lines L1b, L2b, and L3b inclined by 45 ° and L2a and L3a are provided.

  The image reading apparatus of the present embodiment includes a process adjustment mode for adjusting the reading range and the delay amount in the main scanning direction of each CIS and determining the reading range and the delay amount necessary for proper inlining. In addition, after adjustment in the adjustment mode (after correction described later) is applied, a document reading mode for reading a document is provided, and can be arbitrarily switched to the adjustment mode. The adjustment mode is limited to when the main power supply of the image reading apparatus is turned on (at most several times a day), and need not be performed every time the document is read.

  In the adjustment mode, CIS1 performs reading after CIS2 and CIS3, so that no delay in the sub-scanning direction is performed. Further, the CIS1 reading range is not changed in the adjustment in the main scanning direction. By adjusting the reading ranges of CIS2 and CIS3, the section length of each overlapping reading range is adjusted to the target value. After adjusting CIS2 (reading range, delay amount), adjust CIS3.

  The reading range in the adjustment mode uses the execution reading range of each CIS. Further, the reading range in the document image reading mode uses the document reading range including the overlapping reading ranges of CIS1, CIS2, and CIS3. The image signal in the overlapping reading range is corrected (a method will be described later) and incorporated into the image signal that has been made into one line.

  FIG. 3 shows an image signal when a thin line of the contact glass is read in the adjustment mode. The main scanning sequence starts at the left end P1a of the CIS1 execution reading range, reaches the right end P1b of the CIS1 execution reading range, jumps to the left end P2a of the CIS2 execution reading range, and reaches the right end P2b of the CIS2 execution reading range. At the same time, it jumps to the left end P3a of the CIS3 execution reading range, reaches the right end P3b of the CIS3 execution reading range, forms one main scanning period, and repeats scanning.

  The waveform of group A corresponds to the connection between CIS1 and CIS2, and group B corresponds to CIS2 and CIS3. Valley peak waveforms 1b, 1a, 2b, 2a, 3b, 3a, 4a and 4b correspond to signals obtained by reading the thin lines L1b, L1a, L2b, L2a, L3b, L3a, L4a and L4b on the contact glass 5.

  In the waveform of the A group, the interval SL1 corresponds to the thin line L1a and the thin line L2a, and represents the positional relationship between the CIS1 and CIS2 in the main scanning direction. When the relative position in the main scanning direction of CIS1 and CIS2 (that is, the length of the overlapping section) changes, the interval SL1 changes. The interval SG1 also changes. Even if CIS1 and CIS2 move in the sub-scanning direction, the interval SL1 does not change. However, since the thin lines L1b and L2b are inclined, the interval SG1 changes.

  Similarly, in the waveform of group B, the interval SL2 corresponds to the thin line L3a and the thin line L4a, and represents the positional relationship between the CIS2 and CIS3 in the main scanning direction. When the relative position in the main scanning direction of CIS2 and CIS3 (that is, the length of the overlapping section) changes, the interval SL2 changes. The interval SG2 also changes. Even if CIS1 and CIS2 move in the sub-scanning direction, the interval SL1 does not change. However, since the thin lines L1b and L2b are inclined, the interval SG2 changes.

The adjustment method will be described. The adjustment of the joints of CIS1, CIS2, and CIS3 is to make each of the sections SL1, SG1, SL2, and SG2 coincide with preset constants K1, K2, and K3 (adjustment target values).
SL1 ＝ SL2 ＝ K1 ・ ・ ・ Formula (1)
SG1 ＝ K2 ・ ・ ・ Formula (2)
SG2 ＝ K3 ・ ・ ・ Formula (3)

  Depending on the configuration, adjustment target values of SL1, SG1, SL2, and SG2 need to be set individually. Here, there is no problem even if the valley peak waveforms 1b, 1a, 2b, 2a, 3b, 3a, 4a, and 4b are moved left and right on the time axis without changing the relative relationship of the respective positions. CIS1, CIS2, and CIS3 do not change relative positions over time or due to environmental fluctuations, etc., and there are only a few dimensions that cause position fluctuations. If the relative position changes, even a slight problem will be a problem on the read image.

  As values corresponding to the adjustment target values that are the preconditions for adjustment, the overlapping section of CIS1 and CIS2 in the main scanning direction is 2.709 mm (for 64 pixels), and the interval in the sub-scanning direction is 25.400 mm (for 600 pixels). The overlap section between CIS1 and CIS2 is 2.709 mm (64 pixels), and the adjustment in the sub-scanning direction is 23.368 mm (552 pixels, that is, 552 lines). The interval target value in the sub-scanning direction is a value that minimizes the level difference of the joint in the read (or copy) image of the test chart on which the straight line in the main scanning direction is drawn by adjusting the delay amount of CIS2 or CIS3. It may be a target value. Note that the time axis in FIG. 3 corresponds to the pixel interval, and thus the dimensions on each mechanism.

  In the adjustment, first, the reading start position P2a of CIS2 (the left end of the execution reading range) is adjusted in order to set SL1 in the relationship between CIS1 and CIS2 to the target value K1. Since the effective reading length is constant, when the reading start position P2a is determined, the reading end position P2b of the effective reading range is naturally determined. SL1 is matched with K1 by adjusting the reading start position P2a. Thereafter, the interval between the CIS1 and CIS2 in the sub-scanning direction is adjusted by adjusting the image signal delay amount of the CIS2, and the interval SG1 is adjusted to the target value K2. The execution reading range of CIS1 is fixed. The CIS1 signal is not delayed.

  Thereafter, the reading start position P3a of CIS3 (the left end of the execution reading range) is adjusted in order to set SL2 in the relationship between CIS2 and CIS3 to the target value K2. Since the effective reading length is constant, when the reading start position P3a is determined, the reading end position P3b of the effective reading range is naturally determined. SL2 is made to coincide with K1 by adjusting the reading start position P3a. Thereafter, the interval between the CIS2 and CIS3 in the sub-scanning direction is adjusted by adjusting the image signal delay amount of the CIS3, and the interval SG2 is adjusted to the target value K3.

  In the document image reading mode, an image signal in which only one document reading range of each CIS is made into one line is used. The overlap reading range processing (joint signal processing) will be described later. The CIS1 right overlapping reading section on the right side of CIS1 and the CIS2 left overlapping reading section on the left side of CIS2 are configured to perform 64 pixel overlapping reading. The left side of the CIS1 right overlapping reading section has a CIS1 original image reading range other than the overlapping reading, and the right side of the CIS2 left overlapping reading section has a CIS2 original image reading range other than the overlapping reading. There is a CIS1 right position detection section on the right side of the CIS1 right overlap reading section, and a CIS2 left position detection section on the further left side of the CIS2 left overlap reading section.

  Expressions (1), (2), and (3) described above can be established with high accuracy by adjusting the mechanism. However, when the adjustment pitch of the main and sub reading position adjustments is performed in units of pixel intervals (or scanning line intervals) using electrical signals, even after adjustment, the deviation of the reading position is a maximum of 1 pixel interval (or scanning line interval). / 2 (21.1 μm for 600 DPI).

  In the equation (1), when the section SL1 is larger than the target value K1, it indicates that each pixel including the overlapping reading range of CIS2 is shifted to the left with respect to the target (on the vertical line). If it is small, it indicates that it is shifted to the right. In the equation (1), when the section SL2 is larger than the target value K1, it indicates that each pixel including the overlapping reading range of the CIS3 is shifted to the left with respect to the target (on the vertical line). If it is small, it indicates that it is shifted to the right. Hereinafter, seam correction will be described with reference to FIGS. 4 to 6 on the premise that there is a shift in the main scanning direction (therefore, the maximum value is 1/2 of the pixel interval).

  FIG. 4 illustrates the signal processing at the joint between CIS1 and CIS2. FIG. 5 is an enlarged view of a part of FIG. After completion of reading adjustment in the main scanning direction, as shown in FIG. 4, there is no deviation when a connection (straight line) connecting pixels to be overlapped in the overlapping reading ranges of CIS1 and CIS2 is perpendicular to the scanning line. However, there is usually a deviation as shown in FIG.

  Therefore, after shifting (shifting) the signals so that each pixel in each overlapping reading range is in a vertical relationship (both approaching each other) (correction 1), a reading signal is generated using the signal of the corresponding pixel. (Correction 2). After correction 1, the corresponding pixels are aligned on a (virtual) perpendicular.

  The reading signal is a signal corresponding to the overlapping reading range of the inline reading signal (one scanning line is the original reading range of the reading device). Here, when each image signal shift amount (correction 1) is equal to 1/2 of the shift and has a vertical relationship with respect to the signal, for example, the end pixel of the overlapping reading range and other than the overlapping reading range (overlapping reading range) The spacing between the pixels (adjacent to the range) increases and causes image degradation at the joints. Therefore, when shifting, the shift amount on the side that does not contribute to reading the original image in the reading range adjacent to the overlapping pixel (on the right side of the overlapping reading range in CIS1) is increased, and the side that contributes to the reading (on the left side of the overlapping reading range in CIS1) Make small.

  In the adjacent CIS, the side that does not contribute to the document image reading and the side that contributes are reversed in position, so that correction can be performed by a combination of a pixel with a large shift amount and a small pixel (matching the total shift amount with the shift amount). The shift amount is gradually decreased from the shift amount on the side not contributing to the document image reading to the contributing side. The shift amount changes for each pixel. It should be noted that a plurality of pixels in the overlapping reading range can be shifted by the same amount depending on the apparatus (however, depending on the reading characteristics of the joint).

  The increasing / decreasing direction of the shift amount to be gradually changed is opposite in the adjacent CIS. FIG. 7 illustrates how pixels are shifted in the overlapping reading range of CIS1 and CIS2. In the overlapping range of CIS1, the shift amount increases from left to right. On the other hand, in the overlapping range of adjacent CIS2 (on the left side), the shift amount becomes smaller from left to right.

Here, when the deviation from the perpendicular is Δp, the shift amounts Δp1 (CIS1 side) and Δp2 (CIS2 side) of the corresponding pixels are Δp = Δp1 + Δp2 (4)
And Δp is an error equal to or less than the remaining pixel interval after adjusting the reading range in the main scanning direction, and is constant.

  Δp1 increases from left to right. Although it is possible to change the shift amount for each pixel, if the number of pixels in the overlapping reading range is large, the pixels in the overlapping range (64 pixels in this example) are grouped into a plurality of pixels as described above, and each group Are shifted by the same amount of pixels in the same group. In this example, there are 8 sets (64/8). Note that the eight sets of shift amounts change gradually. The number of pixels K (64) in the overlapping reading range on the CIS1 side is equally divided into 8, and the shift amounts of the equally divided 8 pixels are the same.

  The level of the pixel signal after the shift is determined by a convolution method. Δp2 becomes smaller from left to right. Similarly, the number K (64) of pixels in the overlapping reading range on the CIS2 side is equally divided into eight, and the shift amounts of the equally divided eight pixels are the same. The level of the pixel signal D2_1 (n) after the shift is determined by a convolution method.

FIG. 8 and Table 1 show that the signal level σ of the pixel n ′ of correction 1 is set to four neighboring pixels (n−1, n, n + 1, n + 2 pixel levels S _i−1 , S _i , S _{i + 1} , The explanation determined from S _{i + 2} ) is described. The pixel n ′ is a pixel shifted to the right by r in order to arrange on the perpendicular line. The signal level σ of the pixel is obtained from Equation 1 and Table 1. r is a value normalized by the pixel interval.

This is obtained from the residual error (cannot be adjusted by the adjustment pitch) by adjusting the reading range in the main scanning direction. The reading range is adjusted from SL1 and SL2 in FIG. 3, but a residual error occurs because the adjustment pitch is the pixel interval. The value r closest to the value e obtained by dividing the residual error by the pixel interval is selected, and each value is used in Equation 1. Then, using the image signal D1_2 (n) of the CIS1 right overlapping reading section and the image signal D2_2 (n) of the CIS2 left overlapping reading section, the image signal D12 (n) of the CIS1, CIS2 connecting correction section is calculated using the following equation: Generate.
D12 (n) = (k−n) / k × D1_2 (k−n) + n / k × D2_2 (n) (5)
n = 1,2 ... K

An example in which the number of pixels in the CIS1 and CIS2 overlapping reading overlap section is 64 pixels each and grouped by 8 pixels (therefore, the coefficient for the signal is also common for each group) is described below.

When 1 ≦ n ≦ 8 D12 (n) = 7/8 D1-2 (n) + 1 / 8D2-1 (n) (6)
When 9 ≦ n ≦ 16 D12 (n) = 6/8 D1-2 (n) + 2 / 8D2-1 (n) ・ ・ ・ Equation (7)
When 17 ≦ n ≦ 24 D12 (n) = 5/8 D1-2 (n) + 3 / 8D2-1 (n) ・ ・ ・ Equation (8)
When 25 ≦ n ≦ 32 D12 (n) = 4/8 D1-2 (n) + 4 / 8D2-1 (n) ・ ・ ・ Equation (9)
In case of 33 ≦ n ≦ 40 D12 (n) = 4/8 D1-2 (n) + 4 / 8D2-1 (n) ・ ・ ・ Equation (10)
When 41 ≦ n ≦ 48 D12 (n) = 3/8 D1-2 (n) + 5 / 8D2-1 (n) ・ ・ ・ Equation (11)
When 49 ≦ n ≦ 56 D12 (n) = 2/8 D1-2 (n) + 6 / 8D2-1 (n) ・ ・ ・ Equation (12)
When 57 ≦ n ≦ 64 D12 (n) = 1/8 D1-2 (n) + 7 / 8D2-1 (n) ・ ・ ・ Equation (13)

  The image signal D12 (n), the image signal D1-1 of the CIS1 original image reading range other than the duplicate reading, and the image signal D2-3 of the CIS2 original image reading range other than the duplicate reading are 1 near the joint of CIS1 and CIS2. The lined read signal is shown in the lower part of FIG. By smoothly changing the image signal on the right side of CIS1 to the level of the image signal on the left side of CIS2, we aim to make it inconspicuous on the image even if there is a characteristic difference between CIS1 and CIS2 or a slight misalignment. . The signal processing at the joint between CIS2 and CIS3 is performed in the same manner as the signal processing at the joint between CIS1 and CIS2.

Using the image signal D2_4 (n) in the CIS2 right overlap reading section and the image signal D3_2 (n) in the CIS3 left overlap reading section, the image signal D23 (n) in the CIS1, CIS2 joint correction section is generated using the following formula To do.

D23 (n) = (k−n) / k × D2_4 (k−n) + n / k × D3_2 (n) (14)
n: 1, 2 ... K

  By the correction method described above, even if the CIS joint is shifted in the main scanning direction, a read image in which the joint cannot be recognized can be obtained. However, if the CIS shifts in the sub-scanning direction, for example, the read image of thin horizontal thin lines may have a portion corresponding to the CIS joint (overlapping reading range), or the density may become light and difficult to read. . If the CIS is misaligned, when a thin horizontal thin line is read on one side (for example, the right overlapping reading range for CIS1) in the overlapping reading range, the other (for example, the left overlapping reading range for CIS2) is white (thin) This is because there is a case where reading is performed in a place where there is no horizontal thin line. The reading signals corresponding to the overlapping reading range are shown in Equations (5) and (6). However, when CIS is shifted in the sub-scanning direction, it is larger than when D12 (n) and D23 (n) are not shifted. As a result, the image becomes thin (image signal large white) and blurring (missing image information) occurs. It is assumed that the line delay in the CIS sub-scanning direction (in this case, the CIS2 signal is delayed) is performed.

  FIG. 9 is a diagram for explaining prevention of blur due to missing image information. The image arrangement is the same as in FIG.

Signals corrected in the main scanning direction (CIS1 correction and CIS2 correction lines) are used (solid arrows), and the equations (5) and (6) are not used for the joints. CIS1 image signal is used as it is up to n of CIS1 overlap reading range (center of overlap reading range), then CIS2 n + 1 is used, then CIS2 image signal to the center of CIS2 and CIS3 overlap reading range Use as is. The CIS3 image signal is used as it is on the right side of the center of the overlapping reading range of CIS2 and CIS3. This is the same as that obtained by changing the coefficients of the equations (6) to (13) as follows.

When 1 ≦ n ≦ 8 D12 (n) = 8/8 D1-2 (n) + 0 / 8D2-1 (n) ・ ・ ・ Expression (15)
When 9 ≦ n ≦ 16 D12 (n) = 8/8 D1-2 (n) + 0 / 8D2-1 (n) ・ ・ ・ Equation (16)
When 17 ≦ n ≦ 24 D12 (n) = 8/8 D1-2 (n) + 0 / 8D2-1 (n) ・ ・ ・ Equation (17)
When 25 ≦ n ≦ 32 D12 (n) = 8/8 D1-2 (n) + 0 / 8D2-1 (n) ・ ・ ・ Equation (18)
In case of 33 ≦ n ≦ 40 D12 (n) = 0/8 D1-2 (n) + 8 / 8D2-1 (n) ・ ・ ・ Equation (19)
When 41 ≦ n ≦ 48 D12 (n) = 0/8 D1-2 (n) + 8 / 8D2-1 (n) ・ ・ ・ Equation (20)
When 49 ≦ n ≦ 56 D12 (n) = 0/8 D1-2 (n) + 8 / 8D2-1 (n) ・ ・ ・ Equation (21)
When 57 ≦ n ≦ 64 D12 (n) = 0/8 D1-2 (n) + 8 / 8D2-1 (n) ・ ・ ・ Equation (22)

  In equations (6) to (13), the coefficients are gradually changed from 1/8 to 7/8, but in equations (15) to (22), 0 or 8/8 (= 1) is used. Various values can be used as the coefficient, but the coefficient is determined to have a small image quality deterioration at the joint (overlapping reading range) in consideration of the reading characteristics of the apparatus and the means for realizing the signal processing. It depends on the target image quality.

  In the above description, the connection position of the image signal is the center of the overlapping reading range, but it may not be the center. If the connection position is conspicuous on the read image when the connection position is determined, the connection position is changed for each main scanning line (or for every several lines). Also, a simple method without signal correction in the main scanning direction (CIS 1 element, CIS 2 element rows) is possible. 9 indicates that there is no signal correction in the main scanning direction (shown between CIS1 and CIS2; the same applies between CIS2 and CIS3), and the solid line arrow indicates the case where signal correction in the main scanning direction is performed. The connection position is variable.

By the above, the system which combined each method is realizable. The presence or absence of signal correction in the main scanning direction does not make sense to be mounted together (in the following description example, “with main scanning correction” is changed to “without main scanning correction”).

Method 1: With main scanning correction + With sub scanning correction (Coefficient fixed during sub scanning correction)
Method 2: With main scanning correction + With sub scanning correction (Coefficient for sub scanning correction is changed for each scanning line)
Method 3: Main scanning correction + Sub scanning correction (Sub scanning correction coefficient is changed for each of multiple scanning lines)
Method 4: With main scanning correction + Without sub-scanning correction (fixed connection position)
Method 5: With main scanning correction + without sub-scanning correction (change the connection position for each scanning line within the overlapping reading range)
Method 6: With main scanning correction + without sub-scanning correction (change the connection position to multiple scanning lines within the overlapping reading range)
Method 7: Combination of a plurality of methods in the above methods 1 to 6

  Here, the number of pixels at the connection position (moving to the left and right within the overlapping reading range) and the number of the plurality of scanning lines can be set in advance (in an adjustment mode or the like) from an operation unit (not shown).

  Each method is selected according to the original to be read and the purpose of the user. Methods 1, 2, and 3 give priority to image naturalness (small jaggy), and methods 4, 5, and 6 give priority to resolution. Method 7 (which selects at least one of methods 1, 2, 3 and at least one of methods 4, 5, 6) balances the naturalness and resolution of the image.

By implementing this embodiment, the following effects can be obtained. In other words, in the image reading apparatus in which the CISs are arranged in a staggered manner, in the reading range adjustment of each CIS in the main scanning direction, even if an adjustment error occurs, pixels in the overlapping reading range are aligned in the sub-scanning direction by pseudo pixels. As a result, correction based on alignment is possible with high accuracy. In addition, pseudo-pixels are created after coarse adjustment at the pixel interval pitch corresponding to the DPI of the image reading device, so that pseudo-pixels can be created more accurately, and as a result, there is no sense of incongruity (at joints and other parts) (The difference is not visually noticeable).
Even when an adjustment error occurs, when the pixels in the overlapping reading range are aligned in the sub-scanning direction by using pseudo pixels, the sum of the movement amounts of the two target pseudo pixels from the target pixel is made equal to the adjustment error. Therefore, it is possible to obtain an image quality that is less uncomfortable than the case where the alignment is performed by moving only one side (the difference between the joint and the other portion is not visually noticeable).
Even when an adjustment error occurs, when the pixels in the overlapping reading range are aligned in the sub-scanning direction by the pseudo pixels, the sum of the movement amounts of the two target pseudo pixels from the target pixel is equal to the adjustment error. In addition, the amount of movement of the pseudo pixel from the target pixel is changed between the original reading (each CIS center) side and the opposite side of the overlapping reading range (small on the CIS center side and large on the opposite side) An image quality with no sense of incongruity at the boundary portion between the reading range and another reading range (the difference between the boundary portion and other portions is not visually noticeable) is obtained.

Moreover, by implementing this embodiment, the following effects can be obtained. That is, in the image reading apparatus in which the CISs are arranged in a staggered manner, a plurality of signal corrections are used in combination at the joints of the CISs, so that a read image corresponding to the purpose can be obtained while preventing image deterioration.
Since a new image signal necessary for inlining is created (signal correction) in units of scanning lines, the configuration and design of each process such as correction becomes easy. In addition, in the signal correction at each CIS joint, the correction specification is changed by changing the coefficient, so that a desired specification can be easily obtained. In addition, since each CIS connection is directly connected at the connection position without being affected by the signals of other CIS, it is possible to perform image reading without missing information or blurring.
Further, since the position of the CIS joint is moved, the joint is made inconspicuous, and image reading with high naturalness can be performed.
In addition, since the user can select a seam correction method or the like, desired image reading can be performed while preventing image deterioration.

  The above-described embodiment is the best mode for implementing the image reading apparatus of the present invention, but is not intended to be limited to such an embodiment. Therefore, various modifications can be made to the implementation form without departing from the scope of the present invention.

  Development of a FAX, scanner, MFP (Multi Functional Peripheral), etc., equipped with the image reading apparatus of the present invention is desired.

It is sectional drawing of an image reading apparatus. It is a top view of an image reading apparatus. It is an image signal when a thin line of contact glass is read in the adjustment mode. The state of signal processing at the joint between CIS1 and CIS2 is illustrated. It is the figure which expanded a part of FIG. The connection (straight line) connecting the pixels that should correspond in the overlapping reading range of each CIS is illustrated as being shifted from the scanning line. FIG. 6 illustrates how pixels are shifted in the overlapping reading range of CIS1 and CIS2. The method for determining the signal level σ is illustrated. It is a figure explaining the prevention of the blurring by the omission of image information.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Document insertion sensor 2 Conveyance roller: Front 3 Registration sensor 4 Reading back member 5 Contact glass 6 Conveyance roller: Rear

Claims (7)

Are arranged in a plurality of image sensors staggered, and each of the image sensor, the adjacent image reading apparatus effective reading range of the image sensor and the main scanning direction is arranged to have overlapping effective reading range by a predetermined amount duplicate to ,
A plurality of first thin lines parallel to the sub-scanning direction and a contact glass on which a plurality of second thin lines having an angle with respect to the main scanning direction and the sub-scanning direction are formed;
An acquisition unit that , among the plurality of image sensors, a detection range that is outside the overlapping effective reading range of an adjacent image sensor in a main scanning direction acquires an interval between first signals obtained by reading the first thin line. When,
Among the plurality of image sensors, the detection range of the adjacent image sensor is upstream of the document conveyance direction among the plurality of image sensors based on the interval between the second signals obtained by reading the second thin line. Setting means for setting a delay time when the image sensor on the side combines the data obtained by reading the document;
Delay means for delaying data read by the upstream image sensor based on the delay time when reading a document with the plurality of image sensors;
Based on the interval of the first signal acquired by the acquisition unit , a signal obtained from the overlapping effective reading range of each image sensor, the overlapping reading range corresponding to the inner side in the main scanning direction of each image sensor is obtained. such that said amount of movement of overlapping reading the redundant read range is not part signal read in the near side of the range is smaller than the image sensor end side in the overlapping reading range, gradually changing the movement amount Moving means for moving a signal obtained from the overlapping reading range in the main scanning direction;
A combining unit configured to combine a signal obtained from the overlap reading range of each of adjacent image sensors among the plurality of image sensors and subjected to the delay by the delay unit and the signal moved by the moving unit;
Image reading apparatus, characterized in that it comprises a. The movement in the main scanning direction by the moving means is a movement in which the adjacent image sensors read the same position in the main scanning direction so that two pixels respectively obtained at the same position in the main scanning direction. The image reading apparatus according to claim 1, wherein the image reading apparatus is provided. The image reading apparatus according to claim 1, wherein a total of the movement amounts of one of the adjacent image sensors by the moving unit is equal to a total of the movement amounts of the other. The said synthetic | combination means is provided with the determination means which determines the level of each pixel signal based on the signal output from two said adjacent image sensors, The any one of Claim 1 thru | or 3 characterized by the above-mentioned. Image reading device. A signal output from the combining unit based on an output from each image sensor is used from a portion that is not the overlapping reading range to a predetermined position in the overlapping reading range of each of the image sensors. The image reading apparatus according to any one of claims 1 to 3.   The image reading apparatus according to claim 5, wherein the predetermined position is changeable.   An image forming apparatus comprising the image reading apparatus according to claim 1.

Images