JP3299873B2 - Image processing apparatus and image processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and a method for automatically extracting an area of each sheet from an image including a plurality of standard sheets, for example, reading a plurality of business cards at once. , Business card area is automatically extracted and character recognition
Used to create an address book.

[0002]

2. Description of the Related Art When a document is read as an image and is automatically recognized, only one sheet is present in the read image, and after checking the range and inclination of the sheet, recognition within the sheet is performed. It is common. For example, a commercially available business card reader reads a business card one by one as an image and then performs a recognition process.

[0003]

SUMMARY OF THE INVENTION According to the present invention, a plurality of fixed-size sheets are read at once as an image, an area of each sheet is automatically and quickly extracted, and data of the extracted area is subjected to character recognition processing. In this way, characters printed on standard paper such as a plurality of business cards are recognized at a time.

[0004]

An image processing apparatus according to the present invention reads and inputs a plurality of fixed-size sheets as images, extracts an area for each sheet, and performs recognition processing on the extracted area. An image processing apparatus having a function, for the input image, scans the input image in the vertical and horizontal directions and extracts a white pixel to obtain a peripheral distribution. By dividing a continuous area where white pixels in the vertical direction and the horizontal direction are present as an effective section, and determining each of the divided sections as an image area where a white pixel exists in the input image. The image processing apparatus further includes an area extraction control unit that extracts an area for each of the sheets . 1) The area extraction control unit determines whether each of the divided sections
The start coordinates and end coordinates are extracted from the
Calculate the section length from the target and set the section length to a predetermined reference value.
Compare, if less than the reference value, the white pixel in the section
By deleting the data of
It is determined that there is no element. 2) The region extraction control means determines whether each of the divided sections
The start and end coordinates are extracted from the
From the start coordinate and end coordinate, the length between both sections and the total of both sections
The length is determined, and each determined length is a predetermined standard.
Value, and if smaller than the reference value,
Both sections are set so that the data of white pixels in
By integrating, the area consisting of the section has white pixels.
Area. 3) The area extraction control means determines whether each of the divided sections
The start coordinates and end coordinates are extracted from the
The section length is obtained from the target, and the section length is the predetermined fixed form.
Incorrect reading error that has been set to a length that is a natural number times the length of the paper
If the section length is included in the range of the length including the difference,
Divided into sections of a predetermined form length multiplied by a natural number
By doing so, the area consisting of the section is
Area. It is characterized by the following.

[0005] In the image processing apparatus according to the present invention, in the above-mentioned image processing apparatus, the area extraction control means may determine a region where a white pixel defined from each of the divided sections exists by a plurality of times in the region. Is performed by linear scanning.

An image processing method according to the present invention has a function of reading and inputting a plurality of fixed-size sheets as images, extracting an area for each sheet, and recognizing the extracted area. An image processing method of the apparatus, wherein the input image is scanned in a vertical direction and a horizontal direction in the input image, and a peripheral distribution is obtained by extracting white pixels.
Using this peripheral distribution, a continuous area where white pixels in the vertical and horizontal directions are present is divided into sections as effective sections, and each of the divided sections is used as an image area where white pixels exist in the input image. have a region extraction step of extracting an area of each of said piece by determining, 1) the area extracting step, whether the respective divided sections
The start coordinates and end coordinates are extracted from the
Calculate the section length from the target and set the section length to a predetermined reference value.
Compare, if less than the reference value, the white pixel in the section
By deleting the data of
It is determined that there is no element. 2) The region extraction step is performed for each of the divided sections.
The start and end coordinates are extracted from the
From the start coordinate and end coordinate, the length between both sections and the total of both sections
The length is determined, and each determined length is a predetermined standard.
Value, and if smaller than the reference value,
Both sections are set so that the data of white pixels in
By integrating, the area consisting of the section has white pixels.
Area. 3) The region extraction step is performed for each of the divided sections.
The start coordinates and end coordinates are extracted from the
The section length is obtained from the target, and the section length is the predetermined fixed form.
Incorrect reading error that has been set to a length that is a natural number times the length of the paper
If the section length is included in the range of the length including the difference,
Divided into sections of a predetermined form length multiplied by a natural number
By doing so, the area consisting of the section is
Area. A method characterized by the following. In the image processing method according to the present invention, in the above-described image processing method, the area extraction step may determine a region where a white pixel defined from each of the divided sections exists by performing a plurality of linear scans in the region. The method is characterized in that the method is performed by:

(Operation) In the present invention, the peripheral distribution of an input image is used
Divide vertically and horizontally, and calculate the length of the section
Combine or delete sections by comparing predetermined values
And white pixels in the area defined by the vertical and horizontal sections
Is to extract the area where exists, and by integrating or deleting the vertical and horizontal sections based on the predetermined value, the area of the appropriate size is extracted, and the noise area is extracted. Can be ignored.

Further, in the present invention, the length of the obtained section is
If the value is an integral multiple of the predetermined value,
Is divided into multiple sections with the length of the value
Therefore, even when the paper approaches and the area of the paper cannot be separated using the peripheral distribution, it is possible to extract the area of each sheet. Further, in the present invention,
Judgment of the presence of white pixels in the area
Since the inspection is performed by inspection, it is possible to perform high-speed processing by performing a minimum necessary scan in order to confirm whether a sheet actually exists in an area obtained using the peripheral distribution.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an image processing apparatus for extracting a plurality of regions according to the present invention will be described below with reference to FIGS.
2 will be described.

FIG. 1 is a functional block diagram showing the configuration of an embodiment system to which the present invention is applied. S is a scanner for taking in an image, M1 is a memory for input images, M2 and M3 are memories for peripheral distribution, M4 and M5 are memories for section coordinates, R
Is a program ROM, and P is a control unit that controls the flow of processing according to the program in R.

FIG. 2 to FIG. 10 are flowcharts showing the flow of the processing of the embodiment. Note that the standard paper width is XP and the standard paper height is YP.

A region is extracted from an image as follows. [Step (1) in FIG. 2] The image read by the scanner S is transferred to the input image memory M1. An image is stored in an image (horizontal: X pixels, vertical: Y pixels) in order from the upper left pixel to the lower right pixel in the image. "0" is 1by
te is stored in the memory. That is, the coordinates (x,
The pixel value of y) is stored at the (x + X × y) -th address of the memory in units of 1 byte. The memory capacity of one image requires (X × Y) bytes.

[Step (2) in FIG. 2] Based on the image stored in the input image memory M1, the control unit P obtains the peripheral distribution in the image and stores the values in the peripheral distribution memories M2 and M3. I do. The peripheral distribution is obtained by scanning the area in the vertical or horizontal direction and counting the number of white pixels. The peripheral distribution PPv in the vertical direction and the peripheral distribution PPh in the horizontal direction in the image M1 are obtained by dividing i by the horizontal direction. Direction coordinates,
When j is a coordinate in the vertical direction, the following formula (1),
It is expressed as Equation (2).

[0014]

(Equation 1)

At this time, the section where PPv (i) = 0 is a portion where there is no area continuously obtained in the vertical direction, and PPh (j) = 0
Section corresponds to a portion where there is no area to be continuously obtained in the horizontal direction. The storage method is such that, for the peripheral distribution by vertical projection, PPv (i) = i = 0 to (number of horizontal pixels−1).
If 0, “0” is stored in PPv (i) ≠ 0, and “1” is sequentially stored in the peripheral distribution memory M2 in units of 1 byte.
Regarding the marginal distribution by the horizontal projection, “0” when PPh (j) = 0 in j: 0 (the number of vertical pixels−1).
And if PPh (j) ≠ 0, “1” is stored in the peripheral distribution memory M
3 are sequentially stored in units of 1 byte. The memory capacity is required for M2: the number of horizontal pixels, and for M3: the number of vertical pixels, byte.

[Step (3) in FIGS. 2 and 3] Controller P
Scans the inside of the peripheral distribution memory M2, finds a section where M2 (i) = 1, and stores the section start coordinate value / end coordinate value in the section coordinate memory M4 (FIG. 11). Similarly, the control unit P scans the inside of the peripheral distribution memory M3, finds a section where M3 (j) = 1, and stores the section start coordinate value / end coordinate value in the section coordinate memory M5.

[Step (4) in FIG. 4] The control section P retrieves the section start / end coordinate values from the section coordinate memory M4, finds the section length, and sets the section length to a predetermined value Th.
If it is less than sx (= the minimum value of the horizontal distance between sheets), the section data is deleted. Also, the section start / end coordinate values are extracted from the section coordinate memory M5, the section length is determined, and if the section length is less than a predetermined value Thsy (= the minimum value of the vertical distance between sheets), the section is determined. Delete data. Thereby, the noise component can be removed from the data.

[Step (5) in FIGS. 5 and 6] Control unit P
Extracts the section start / end coordinate values of two adjacent sections from the section coordinate memory M4, and sets the distance between the two sections to Th.
If sx or less and the section length when both sections are integrated are equal to or less than a predetermined value Thlx (= XP + error), both sections are integrated into one section data. The section start / end coordinate values of two adjacent sections are taken out from the section coordinate memory M5, the distance between the two sections is equal to or less than Thsy, and the section length when both sections are integrated is a predetermined value Thly. In the case of (= YP + error) or less, both sections are integrated into one section data. As a result, even if one area is divided into a plurality in the peripheral distribution data,
It can be extracted as one area.

[Step (6) in FIG. 7] The control unit P takes out the section start / end coordinate values from the section coordinate memory M4, finds the section length, and sets a predetermined value to Thn (= reading error). Then, (XP × n
-Thn) ≦ District between length ≦ (X P × n + Thn ) (n: if a natural number) is satisfied, aliquoted its section into n.
That is, [start coordinate value + (i−1) × (end coordinate value−
Start coordinate value) / n, start coordinate value + i x (end coordinate value-
The n pieces of section data of (start coordinate value) / n] (i: 1 to n) are stored in M4 instead of the original data. The section start / end coordinate values are extracted from the section coordinate memory M5, the section length is obtained, and (YP × n−Thn) ≦ section length ≦ (YP ×
If (n + Thn) (n: natural number) holds, the section is equally divided into n sections. Thereby, even when a plurality of regions are one in the marginal distribution data, they can be separated and extracted.

[Step (7) in FIGS. 8 and 9]
Fetches the section start / end coordinate values from the section coordinate memory M4, finds the section length, and when the section length is smaller than a predetermined value Thmx (= XP-error) or Thlx
If it is larger, the section data is deleted. Also, the section start / end coordinate values are extracted from the section coordinate memory M5, the section length is obtained, and the section length is set to a predetermined value Th.
If it is smaller than my (= YP-error) or larger than Thly, the section data is deleted. Thus, when a sheet other than the standard sheet is used, it can be removed from the data.

[Step (8) in FIG. 10] The control unit P
The section start / end coordinate values in the horizontal direction are taken out from the section coordinate memory M4, and the section start / end coordinate values in the vertical direction are taken out one by one from the section coordinate memory M5. Check if it exists. Specifically, the controller P scans the input image memory M1 linearly from the upper left corner to the lower right corner of the rectangle to check the number of white pixels.
5) In the above case, it is assumed that the rectangular area is the area of the sheet to be obtained. When the number is less than Thc, the number of white pixels is further examined by linear scanning rightward from the middle point of the left end of the rectangle, and when the number is greater than Thc, the rectangular area is assumed to be a paper area to be obtained ( (FIG. 12). If the number is less than Thc, it is assumed that there is no paper area to be found in the rectangular area. It is possible to determine whether or not a sheet exists at a higher speed than when scanning the entire area. For the scanning direction, it is also possible to reduce the detection omission of the area by combining various scans such as the lower right direction from the upper right of the rectangle, the upper left direction from the lower right of the rectangle, the upper right direction from the lower left of the rectangle, the lower direction from the middle point of the upper end of the rectangle. It is.

The image data present in the plurality of rectangular areas extracted as described above is introduced into a character recognition unit (not shown), and character recognition is executed for each area, and the recognition result is stored in an address book memory (not shown). And so on.

[0023]

As described above, according to the present invention, a plurality of fixed-size sheets are read at once as an image, an area of each sheet is automatically and rapidly extracted, and data of the extracted area is subjected to character recognition processing. it is, that Do is possible to print characters in standard size paper, such as a plurality name card recognized processed at one time.

[0024] Thus, <br/> in the image processing apparatus according to the present invention, based on the predetermined threshold value, by integrating or remove the vertical and horizontal direction of the section, a suitably sized Regions can be extracted and noise regions can be ignored.

[0025] In the image processing apparatus according to the present invention, the paper approaches, in the case that can not be separated one by one area of the plurality of sheets using the marginal distribution also, on the basis of the magnitude of the predetermined paper Thus, an area of each sheet can be extracted.

[0026] In the image processing apparatus according to the present invention further fast to see if in fact the paper in a region determined using the marginal distribution is present, by performing the required minimum linear scan, handle Can be.

[Brief description of the drawings]

FIG. 1 is a functional block diagram showing a configuration of a system according to an embodiment to which the present invention is applied.

FIG. 2 is a flowchart showing an operation of the system according to the embodiment to which the present invention is applied.

FIG. 3 is a flowchart illustrating an operation of the system according to the embodiment to which the present invention is applied;

FIG. 4 is a flowchart showing an operation of the system according to the embodiment to which the present invention is applied.

FIG. 5 is a flowchart illustrating an operation of the system according to the embodiment to which the present invention is applied;

FIG. 6 is a flowchart illustrating an operation of the system according to the embodiment to which the present invention is applied;

FIG. 7 is a flowchart showing an operation of the system according to the embodiment to which the present invention is applied.

FIG. 8 is a flowchart showing an operation of the system according to the embodiment to which the present invention is applied.

FIG. 9 is a flowchart showing an operation of the system according to the embodiment to which the present invention is applied.

FIG. 10 is a flowchart showing the operation of the system according to the embodiment to which the present invention is applied.

FIG. 11 is a diagram showing an embodiment of a method of storing a value in a section coordinate memory.

FIG. 12 is a diagram showing an embodiment of a method for scanning a rectangular area.

[Explanation of symbols]

 S scanner M1 Input image memory M2, M3 Peripheral distribution memory M4, M5 Section coordinate memory R Program ROM P control unit

Claims (8)

(57) [Claims] 1. An image processing apparatus having a function of reading and inputting a plurality of fixed-size sheets as images, extracting an area for each sheet, and performing a recognition process on the extracted area. The input image is scanned in the vertical direction and the horizontal direction in the input image, and a peripheral distribution is obtained by extracting white pixels. White pixels in the vertical direction and the horizontal direction exist using the peripheral distribution. Area extraction control for extracting a region for each sheet by dividing a continuous region as an effective region and determining each of the divided regions as an image region where a white pixel exists in the input image. Means, and the area extraction control means starts from each of the divided sections.
Extract the start and end coordinates, and calculate the start and end coordinates
From section and compare the section length with a predetermined reference value
If it is smaller than the reference value, the white pixel data
By deleting the data, the white pixels
An image processing apparatus for determining an area that does not exist . 2. A plurality of standard papers are read as images.
After input, an area for each sheet is extracted, and the extracted area is extracted.
An image processing device with a function of recognizing the area
Te, to the input image, the vertical direction in the input image
And by scanning in the horizontal direction and extracting white pixels, the marginal distribution
And use this marginal distribution to create vertical and horizontal white images.
The continuous area where the element exists
Is a white pixel in the input image.
Is determined as an image area where
An area extracting control means for extracting a region of the region extraction control means, opening from the respective divided sections
Extracts the start and end coordinates, and the start position for two adjacent sections
From the mark and the end coordinates, the length between both sections and the total length of both sections
Calculated, and the obtained lengths are determined by a predetermined reference value.
Each is compared, and if smaller than the reference value,
Combine both sections so that white pixel data becomes one section
By doing so, the area consisting of the section is
Images processor characterized by determining the frequency. 3. A plurality of standard papers are read as images.
After input , an area for each sheet is extracted, and the extracted area is extracted.
An image processing device with a function of recognizing the area
Te, to the input image, the vertical direction in the input image
And by scanning in the horizontal direction and extracting white pixels, the marginal distribution
And use this marginal distribution to create vertical and horizontal white images.
The continuous area where the element exists
Is a white pixel in the input image.
Is determined as an image area where
An area extracting control means for extracting a region of the region extraction control means, opening from the respective divided sections
Extract the start and end coordinates, and calculate the start and end coordinates
Is calculated from the section length, and the section length is determined by the predetermined fixed form paper.
Is a natural number times the length of
If the section length is included in the range of the included
Divide into sections of a predetermined standard paper length that is several times larger
Thus, the region including the section is defined as a region where white pixels exist.
Images processor wherein the determining. Wherein said area extraction control means being the divided
The area where the white pixel defined from each section exists
The image processing apparatus according to any one of claims 1 to 3, wherein the performing by multiple linear scan within the region. 5. A plurality of standard papers are read as images.
After input, an area for each sheet is extracted, and the extracted area is extracted.
Image processing of an image processing apparatus having a function of recognizing the area
A physical method, with respect to the input image, the vertical direction in the input image
And by scanning in the horizontal direction and extracting white pixels, the marginal distribution
And use this marginal distribution to create vertical and horizontal white images.
The continuous area where the element exists
Is a white pixel in the input image.
Is determined as an image area where
And a region extracting step of extracting a region from the divided sections.
Extract the start and end coordinates, and calculate the start and end coordinates
From section and compare the section length with a predetermined reference value
If it is smaller than the reference value, the white pixel data
By deleting the data, the white pixels
An image processing method characterized by determining that a region does not exist
Law . 6. An image processing method of an image processing apparatus having a function of reading and inputting a plurality of fixed-size sheets as images, extracting a region for each sheet, and performing a recognition process on the extracted region. The input image is scanned in the vertical and horizontal directions in the input image, and a peripheral distribution is obtained by extracting white pixels. The continuous area where pixels are present is divided into sections as effective sections, and each of the divided sections is determined as an image area where white pixels are present in the input image, thereby extracting the area for each of the individual sheets. have a region extraction step of, the area extracting step, open from the respective divided sections
Extracts the start and end coordinates, and the start position for two adjacent sections
From the mark and the end coordinates, the length between both sections and the total length of both sections
Calculated, and the obtained lengths are determined by a predetermined reference value.
Each is compared, and if smaller than the reference value,
Combine both sections so that white pixel data becomes one section
By doing so, the area consisting of the section is
An image processing method characterized by determining a region . 7. A plurality of standard papers are read as images.
After input, an area for each sheet is extracted, and the extracted area is extracted.
Image processing of an image processing apparatus having a function of recognizing the area
A physical method, with respect to the input image, the vertical direction in the input image
And by scanning in the horizontal direction and extracting white pixels, the marginal distribution
And use this marginal distribution to create vertical and horizontal white images.
The continuous area where the element exists
Is a white pixel in the input image.
Is determined as an image area where
And a region extracting step of extracting a region from the divided sections.
Extract the start and end coordinates, and calculate the start and end coordinates
Is calculated from the section length, and the section length is determined by the predetermined fixed form paper.
Is a natural number times the length of
If the section length is included in the range of the included
Divide into sections of a predetermined standard paper length that is several times larger
Thus, the region including the section is defined as a region where white pixels exist.
Images processing how to wherein the determining. 8. The method according to claim 8, wherein the step of extracting the area includes
The area where the white pixel defined from each section exists
8. The image processing method according to claim 5, wherein the scanning is performed by a plurality of linear scans in the area .