JP2000187706A - Paper sheets reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To precisely acquire a necessary character on paper sheets by judging the presence or absence of a character area in a specified area which is set, detecting a connected image without a rift and identifying the character from a character string in the connected image. SOLUTION: The image signal of paper sheets P, which is image-picked up by an image pick up part 1, is converted into a multilevel digital signal in an A/D conversion part 2. Then, a binarization part 3 binarizes it. A labeling coordinate extraction part 4 detects the coordinates of connected images without a rift among binarization images. A projective area setting part 6 sets plural projective areas as specified areas to the binarization image. When a character area judgement part 7 judges the presence of a character area in the projective area, a character detection/segmenting part 5 detects and segments a character fitted to a prescribed condition in the connected image. A character string judgment part 8 supplies a character string having the character fitted to the prescribed condition among the detected characters to a normalization/sampling part 9 and it is normalized/sampled. A character identification part 11 identifies the character of the character string with reference with a character dictionary 12.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sheet reader for reading sheets.

[0002]

2. Description of the Related Art When reading characters on a paper sheet, an image is taken on the paper sheet, characters are extracted from an image obtained by the imaging, and the extracted characters are compared with a character pattern stored in a character dictionary in advance. Identify by matching. The type of paper sheet may be determined according to the identification result.

[0003]

The image on the paper sheet may be simple, but when a complicated image such as many characters or pictures is described, only the necessary characters are selected from among them. It becomes very difficult to read. This makes it difficult to determine the type of paper sheet.

[0004] The present invention has been made in view of the above circumstances,
An object of the present invention is to provide a paper sheet reading apparatus capable of accurately capturing necessary characters on a paper sheet irrespective of surrounding image conditions.

Another object of the present invention is to provide a paper sheet reading apparatus capable of accurately detecting the type and the transport state of the paper sheet.

[0006]

According to a first aspect of the present invention, there is provided a sheet reading apparatus, comprising: an image pickup means for picking up an image of a sheet;
Area setting means for setting a specific area for an image picked up by the image pickup means; character area determining means for determining whether a character area exists in the specific area set by the area setting means; A connected image detecting means for detecting a continuous connected image among the images picked up by the means, and a character among the connected images detected by the connected image detecting means when the determination result of the character area determining means is positive. A character string detecting means for detecting a string, and a character identifying means for identifying a character of a character string detected by the character string detecting means.

According to a second aspect of the present invention, there is provided a sheet reading apparatus.
Imaging means for capturing images on both front and back sides of paper sheets,
Area setting means for setting a specific area for both front and back images picked up by the image pickup means, and character area determining means for determining whether a character area exists in each specific area set by the area setting means A connected image detecting means for detecting a continuous connected image among the images picked up by the image pickup means, and a connected image detected by the connected image detecting means when the determination result of the character area determining means is affirmative. The image processing apparatus includes character string detecting means for detecting a character string in an image, and character identifying means for identifying a character of the character string detected by the character string detecting means.

According to a third aspect of the present invention, there is provided a sheet reading apparatus.
In the invention according to claim 1 or 2, when the character string detecting means determines that the character area determining means is affirmative, the connected image detected by the connected image detecting means satisfies a predetermined condition. And character string determining means for determining, as a character string, the presence of a character meeting predetermined conditions among the characters detected by the character detecting means.

According to a fourth aspect of the present invention, there is provided a sheet reader.
In the invention according to claim 1 or claim 2,
A determination unit is provided for determining the type of sheet and the transport state according to the identification result of the character identification unit.

According to a fifth aspect of the present invention, there is provided a sheet reading apparatus.
In the invention according to claim 1 or claim 2,
It has a character dictionary that stores upside-down normal character patterns and reverse character patterns for each character. The character identification means identifies a character by comparing it with a character pattern in a character dictionary.

According to a sixth aspect of the present invention, there is provided a sheet reading apparatus.
Imaging means for capturing an image of a paper sheet, divided area setting means for setting a plurality of divided areas for an image captured by the imaging means, and a character string from each divided area set by the divided area setting means A character string detecting means for detecting a character string, a selecting means for selecting a character string meeting a specific condition from among the character strings detected by the character string detecting means, and identifying a character of the character string selected by the selecting means. And a judging unit for judging the type and the transport state of the sheet according to the identification result of the character identifying unit.

According to a seventh aspect of the present invention, there is provided a sheet reading apparatus.
Imaging means for capturing an image of a paper sheet, divided area setting means for setting a plurality of divided areas for an image captured by the imaging means, and a character string from each divided area set by the divided area setting means Character string detecting means, a selecting means for selecting a character string that meets a specific condition from among the character strings detected by the character string detecting means, A first character dictionary storing a pattern and a first character string that is identified by comparing a character of a character string selected by the selection means with a character pattern in the character dictionary.
Character identification means, character detection means for detecting a specific character among the characters detected by the character detection means in accordance with the direction of the collated character pattern used by the first character identification means, A second character identification unit that identifies the specific character to be detected; and a determination unit that determines a type, a conveyance state, and the like of the paper sheet according to identification results of the first and second character identification units.

[0013] According to an eighth aspect of the present invention, there is provided a sheet reader.
In the invention according to claim 6 or 7, the character string detecting means detects a character from each of the divided areas set by the divided area setting means, and the character string detecting means includes: A character string determination unit that determines, as a character string, the presence of a character that meets a predetermined condition.

According to a ninth aspect of the present invention, there is provided a sheet reading apparatus.
An imaging unit that captures an image of a sheet using infrared light and visible light; and a density difference image corresponding to a difference between the density of the image captured by the infrared light and the density of the image captured by the visible light. Density difference image forming means, connected image detection means for detecting a continuous connected image among the density difference images formed by the density difference image forming means, and characters from the connected image detected by the connected image detection means. A character string region detecting means for detecting a line region, and among the images captured by the visible light, detecting a continuous connected image from an image corresponding to the character string region detected by the character string region detecting means. Connected image detecting means, character detecting means for detecting a character from the connected image detected by the connected image detecting means, and character identifying means for identifying a character detected by the character detecting means; Obtain.

According to a tenth aspect of the present invention, there is provided a sheet reading apparatus, comprising: an image pickup means for picking up an image of a sheet by infrared light and visible light; and a density of the image picked up by the infrared light and the visible light. Density difference image forming means for forming a density difference image corresponding to the difference between the density of a captured image and a connected image for detecting a continuous connected image among the density difference images formed by the density difference image forming means Detecting means, character string area detecting means for detecting a character string area from the connected image detected by the connected image detecting means, and character detecting means for detecting a character from the character string area detected by the character string area detecting means And character identification means for identifying a character detected by the character detection means.

According to an eleventh aspect of the present invention, in the paper sheet reading apparatus according to the ninth or tenth aspect, further, the type and transport state of the paper sheet according to the identification result of the character identification means are further determined. The determination means is provided.

According to a twelfth aspect of the present invention, there is provided a paper sheet reading apparatus, comprising: an image pickup means for picking up an image of a paper sheet; an area setting means for setting a specific area for the image picked up by the image pickup means; A projection pattern obtaining means for obtaining a projection pattern from a specific area set by the area setting means, a measuring means for measuring the height of the projection pattern obtained by the projection pattern obtaining means, and the height measured by the measuring means Is determined to be within a predetermined range, and as a result of this determination, when it is determined to be within the predetermined range, the upper projection pattern is determined to be a character area, The character in the projection pattern is identified based on the result of the determination by the determination means.

According to a thirteenth aspect of the present invention, there is provided a sheet reading apparatus, comprising: an image pickup means for picking up an image of a sheet with infrared light and visible light; Density difference image forming means for forming a density difference image corresponding to the difference between the density of the captured image, and a connected image detecting means for detecting a connected image from the density difference image formed by the density difference image forming means, Measuring means for measuring the height and width of the connected image detected by the connected image detecting means; and determining whether the height and width measured by the measuring means are within a predetermined range. And determining means for determining that the projection pattern is a character area when it is determined that the projection pattern is within a predetermined range, based on the determination result by the determining means, In the connected image To identify the character.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, the premise of the present invention will be described. The present invention is to process a sheet having a plurality of predetermined formats. Information such as the height and coordinates of characters described in advance is stored as format information, and is used in the present apparatus.

[1] Hereinafter, a first embodiment of the present invention will be described with reference to the drawings.

In FIG. 1, P is a paper sheet which is conveyed in a certain direction. An imaging unit 1 is provided at a position facing the conveyance path of the paper sheet P, and one side of the paper sheet P is imaged by the imaging unit 1.

A two-dimensional analog image signal on one side of the sheet P picked up by the image pickup unit 1 is converted into a multi-valued digital image signal by the A / D converter 2. The multi-valued digital image signal is binarized by a binarizing unit 3 and supplied to a labeling coordinate extracting unit 4 as a connected image detecting unit and a projection area setting unit 6 as an area setting unit.

The labeling coordinate extraction unit 4 detects (extracts) the coordinates of the start point and the end point of the continuous image that is not interrupted in the captured image (binary image).

The projection area setting section 6 stores the captured image (2
A plurality of so-called projection areas are set as specific areas for the (valued image). The setting contents are supplied to the character area determination unit 7 which is a projection pattern acquisition unit / measurement unit / judgment unit. The character area determination unit 7 determines whether a character area exists in each projection area set by the projection area setting unit 6. The result of this determination is supplied to the character detection and disconnection section 5 which is a character detection means.

When the result of the determination by the character area determination section 7 is affirmative (there is a character area), the character detection section 5 determines a predetermined one of the linked images (the coordinates of the start point and end point) detected by the labeling coordinate extraction section 4. (Height and width within the allowable range compared to the reference character parameter) are detected and cut out as characters in units of one character. This inspection result is supplied to the character string determination unit 8.

The character string judging unit 8 determines a character which meets a predetermined condition (a character pitch, which is an interval between the characters, is within an allowable range compared with a reference value) among the characters detected by the character checking unit 5. Is determined as a character string. That is, the character detection unit 5 and the character string determination unit 8 constitute a character string detection unit.

The character string determined by the character string determination unit 8 is supplied to a normalization / sampling unit 9, where it is normalized and sampled and supplied to a character identification unit 11.

The character identification unit 11 identifies the character of the character string determined by the character string determination unit 8 by comparing it with a character pattern in the character dictionary 12. The character dictionary 12 stores a normal character pattern and a reverse character pattern that are upside down for each character. Then, the identification result of the character identification unit 11 is supplied to the overall determination unit 10.

Further, another image pickup unit 1 for picking up an image of the other side of the sheet P is provided at a position facing the image pickup unit 1 with the conveyance path of the sheet P interposed therebetween. Then, on the other surface side, the A / D conversion unit 2, the binarization unit 3,
A labeling coordinate extracting unit 4, a character checking unit 5, a projecting region setting unit 6, a character region determining unit 7, a character string determining unit 8, a normalizing / sampling unit 9, and a character identifying unit 11 are provided.

The comprehensive determination unit 10 determines the type and transport state (front / back / upright / inverted) of the paper sheet P according to the identification result of each character identification unit 11.

Next, the operation of the above configuration will be described.

FIG. 2 is a flowchart for explaining the operation, and FIGS. 3, 4, 5, and 6 show examples of the conveyed sheets P.

The paper sheet P shown in FIG. 3 has a front surface facing the image pickup unit 1 as one surface, a back surface facing another image pickup unit 1 as the other surface, and a vertical position in an upright state and a forward direction. The state of being conveyed is shown.

The paper sheet P shown in FIG. 4 has a front surface facing the image pickup unit 1 as one surface, a back surface facing another image pickup unit 1 as the other surface, and an upside-down position in an upright state, and the sheet P is transported in the forward direction. It shows how it is done.

The paper sheet P shown in FIG. 5 has a back surface facing the image pickup unit 1 as one surface, a front surface facing another image pickup unit 1 as the other surface, and a vertical position in the upright position. The state of being conveyed is shown.

The paper sheet P shown in FIG. 6 has a rear surface facing the image pickup unit 1 as one surface, a front surface facing another image pickup unit 1 as the other surface, and a vertical position in which the sheet is conveyed in an upright state. It shows how it is done.

All the sheets P are of the same type, and the image A, the character B, the character string C
, And the image D, character E, and character string F
Is described.

Each image pickup section 1 performs a main scan in a direction perpendicular to the transport direction of the paper sheet P. The sub-scan is performed by the main scanning moving along with the conveyance, and one surface and the other surface of the sheet P are respectively imaged. With this imaging, the images on both sides of the sheet P are read (step 100).

Projection areas are set at two positions with respect to the image picked up from one side of the sheet P (step 10).
1). The projection area is projected in the sub-scanning direction as shown in FIGS. 7, 8, 9, 10, 11, 12, 13, and 14 (step 102) to form a projection pattern (step 102). Step 103).

FIG. 7 is a projection pattern obtained from the projection area at the left position of the paper sheet P in FIG. FIG. 8 is a projection pattern obtained from a projection area on the right side of the sheet P in FIG.

FIG. 9 is a projection pattern obtained from the projection area at the left position of the paper sheet P in FIG. FIG. 10 shows the sheet P of FIG.
Projection pattern obtained from the projection area at the right side of.

FIG. 11 shows a projection pattern obtained from a projection area on the left side of the sheet P in FIG. FIG. 12 is a projection pattern obtained from the projection area at the right position of the paper sheet P in FIG.

FIG. 13 shows a projection pattern obtained from the projection area on the left side of the sheet P in FIG. FIG. 14 is a projection pattern obtained from a projection area on the right side of the paper sheet P in FIG.

The projection pattern obtained in this way has a threshold value th.
(Step 104), and the respective heights (lengths along the main scanning direction) h are measured (Step 105).
It is determined whether or not the measured height h falls within a predetermined range α <h <β corresponding to the character area (step 106), and if the condition of α <h <β is satisfied. If it is (YES in step 106), it is determined that the character area exists (is a character area).

An image picked up from one side of the sheet P is labeled (step 107), and the coordinates of the start point and the end point of the continuous image are extracted from the picked-up image (step 108). ). Then, the height (length along the main scanning direction) and width (length along the sub-scanning direction) of the connected image are obtained from the extracted coordinates of the start point and the end point (step 109).

If it is determined in step 106 that there is a character area, the connected image obtained by the labeling whose height and width are within the allowable range compared with the reference character parameter is used. The character is cut off in units of one character (step 110).

FIG. 15 shows an example of a connected image. [1
Each character of the character set (character string) "234567" is a connected image, the height of the connected image is y, the width is x, and the length between centers of adjacent connected images in the sub-scanning direction. Is the pitch p.

[0048] Of the characters to be cut off, the presence of a character whose character pitch is within an allowable range as compared with the reference value is determined as a character string (step 111). This character string is normalized and sampled (step 112), and the characters of the character string are identified by collation with a character pattern in the character dictionary 12 (step 113).

The same processing as described above is also performed on various sides of the paper sheet P (steps 201 to 213). Then, the character identification results obtained from both sides of the paper sheet P are sent to the comprehensive determination unit 10, which determines the type and the transport state of the paper sheet P.

That is, the type is determined based on the content of the identified character, and the character dictionary 1 used for character identification is used.
The transport state (front / back / erect / inverted) can be determined according to whether the character pattern to be verified in 2 is the upward regular character pattern or the downward inverted character pattern.

As described above, necessary characters on the paper sheet P can be accurately captured irrespective of the surrounding image conditions, and the type and transport state of the paper sheet P can be captured accurately.

[2] The second embodiment will be described.

As shown in FIG. 16, an image pickup section 21 is provided at a position facing the conveyance path of the paper sheet P.

The two-dimensional analog image signal on one side of the sheet P picked up by the image pickup unit 21 is converted by the A / D converter 22 into a multi-valued digital image signal. The multi-valued digital image signal is binarized by the binarization unit 23 and supplied to four divided area setting units 24.

Each divided area setting section 24 sets a plurality of divided areas for a captured image. The setting contents are supplied to the four character detection cutout units 25. Each character detection cutout unit 25 constitutes a character string detection unit together with each character string determination / selection unit 26 at the subsequent stage, and detects a character from each divided area set by the divided area setting unit 24. Specifically, a connected image (coordinates of a start point and an end point) is detected by labeling, and a connected image that satisfies a predetermined condition (height and width is within an allowable range compared to the reference character parameter) is extracted from the connected image. Is detected and cut out in units of one character. The inspection result is supplied to the character string determination / selection unit 26.

Each character string determination / selection unit 26 determines the presence of a character that meets a specific condition among the characters detected by each character detection unit 25 as a character string, and determines a specific condition of the character string. Has a selection function (selection means) for selecting a character string conforming to. Specifically, among the detected characters, the presence or absence of a character that meets a predetermined condition (a character pitch that is a mutual interval between the characters is within an allowable range compared with a reference value) is determined for each region as a character string, Corresponding character string data is selectively output according to whether a character string exists in both of the two areas and whether the number of valid characters in the character string is 7 or 8 characters.

The selected character string is converted by the normalizing / sampling unit 27.
, Where it is normalized and sampled and supplied to the character identification unit 29.

The selected character string is supplied to the east / west / north / north character detection unit 28. The east / west / north / north character detection unit 28 sets a position and an area where “east / west / north / north” is described as a specific character in the character string according to character pattern use data described later supplied from the character identification unit 29. Detect a specific character in the setting area. The detection result is the character identification unit 2
9.

The character identification unit 29 is a character string determination / selection unit 2
The character of the character string selected in 6 is identified by collating with the character pattern in the character dictionary 30. The character dictionary 30 stores a normal character pattern and a reverse character pattern that are upside down for each character. Then, the identification result of the character identification unit 29 is supplied to the overall determination unit 10.

A character pattern indicating whether the direction of the character pattern to be verified in the character dictionary 12 used for character identification in the character identification section 29 is an upward normal character pattern or a downward reverse character pattern. The data is supplied from the character identification unit 29 to the east / west / north / north character detection unit 28.

Further, another image pickup section 21 for picking up an image of the other side of the sheet P is provided at a position facing the image pickup section 21 with the conveyance path of the sheet P interposed therebetween. The A / D conversion unit 22, the binarization unit 23, and the divided area setting units 24 are also provided on the other surface side.
5. Character string determination / selection unit 26, normalization / sampling unit 2
7, an east / west / north / south character detection unit 28 and a character identification unit 29 are provided.

The overall determination section 31 determines the type and the transport state (front / back / upright / inverted) of the paper sheet P according to the identification result of each character identification section 29.

Next, the operation of the above configuration will be described.

FIGS. 17, 18, and 19 are flowcharts for explaining the operation, and FIGS. 20, 21, 22, and 23 show examples of the conveyed paper sheets P. FIG.

The paper sheet P in FIG. 20 is a first type of paper sheet.
The front surface is opposed to the imaging unit 1 as one surface, and the back surface is opposed to another imaging unit 1 as the other surface.

The paper sheet P in FIG. 21 is a first type of paper sheet.
The front surface is opposed to the imaging unit 1 as one surface, and the back surface is opposed to another imaging unit 1 as the other surface.

The paper sheet P in FIG. 22 is a second type of paper sheet.
The front surface is opposed to the imaging unit 1 as one surface, and the back surface is opposed to another imaging unit 1 as the other surface.

The paper sheet P in FIG. 23 is a second type of paper sheet.
The front surface is opposed to the imaging unit 1 as one surface, and the back surface is opposed to another imaging unit 1 as the other surface.

In each of the sheets P, an image A, a specific character B, and a character string C are described on the front surface at a substantially left position, and an image D, a specific character E, and a character string F are described at a substantially right position. I have. However, the specific characters B and E are used for the first type and the second type.
There is a difference in the description of the kanji "East, West, South and North", and further, the description positions are different from each other. Also, the first
The type is a character string C having seven valid characters, and the second type is a character string C having eight valid characters.

Each image pickup section 21 performs a main scan in a direction orthogonal to the transport direction of the paper sheet P. The sub-scan is performed by the main scanning moving along with the conveyance, and one surface and the other surface of the sheet P are respectively imaged. With this imaging, images on both sides of the sheet P are read (step 300).

For an image taken from one side of the sheet P, divided areas Q1, Q2, Q3, Q4 which divide the entire surface into four parts
Are set (steps 301, 401, 501, 601). A connected image (coordinates of a start point and an end point) is detected from these divided areas by labeling, and the height and width, which are the character sizes of the connected image, are measured (steps 302, 402, 502, and 602).

If the measured character size is within an allowable range compared to the reference character parameter, the corresponding character string is cut off in units of one character. If the character pitch is within the allowable range compared to the reference value, the presence of the corresponding character is extracted as a character string for each region (Q1, Q2, Q3, Q4). (Step
303,403,503,603). Further, it is determined whether the number of valid characters in the extracted character string is 7 or 8 (steps 304, 404, 504, 604).

The paper sheet P in the upright state of FIG. 20 and FIG.
The paper sheet P in the inverted state is a character string C, F or a specific character B,
E is not in the diagonally located Q1 · Q4 area, and Q2 · Q3
It is in the area, and the number of valid characters of the character strings C and F is 7 characters. This is proof that the paper sheet P is the first type.

The paper sheet P in the upright state of FIG. 22 and FIG.
The paper sheet P in the inverted state is a character string C, F or a specific character B,
E is in the diagonally located Q1 · Q4 area, not in the Q2 · Q3 area, and the number of effective characters of the character strings C and F is eight. This is a proof that the paper sheet P is a second type.

If it is the first type of sheet P (step 404)
(YES in step 504, YES in step 504), and the character strings extracted from the Q2 and Q3 areas are determined to be character strings C and F, respectively, according to the determination factor of the first type (step 504).
405,505). Then, the characters of the determined character strings C and F are normalized and sampled (steps 406 and 506) and identified (steps 407 and 507).

At the time of this character identification, if the sheet P is of the first type and is erected as shown in FIG.
A regular character pattern in 0 is used (Y in step 408).
ES, YES in step 508). In this case, the character strings C,
Based on the position of F, the positions and areas of specific characters B and E are set as shown in FIG. 24 (steps 409 and 50).
9). The relative positions of the character strings C and F and the specific characters B and E are indicated by arrows in the figure. Then, specific characters B and E are detected from the set position and area (steps 410 and 51).
0) and it is normalized and sampled (steps 411,51)
1) identified (steps 412, 512).

If the sheet P is of the first type and is inverted and conveyed as shown in FIG. 21, the reverse character pattern in the character dictionary 30 is used (NO in step 408, N in step 508).
O). In this case, based on the positions of the character strings C and F, the positions and areas of the specific characters B and E are set as shown in FIG. 25 (steps 413 and 513). Then, specific characters B and E are detected from the set position and area (step
414,514), which are normalized and sampled (step 4
15,515) and are identified (steps 416, 516).

On the other hand, if the paper sheet P is of the second type (YES in step 304a and YES in step 604a), the character string extracted from the Q1 and Q4 areas is determined according to the determination factor of the second type. Character strings C and F are determined respectively (steps 405a and 505a). Then, the determined character strings C and F
Are normalized and sampled (steps 406a, 506).
a) is identified (steps 407a, 507a).

At the time of this character identification, if the sheet P is of the second type and is erected as shown in FIG.
A regular character pattern in 0 is used (Y in step 317).
ES, YES in step 617). In this case, the character strings C,
Based on the position of F, the positions and areas of specific characters B and E are set as shown in FIG. 26 (steps 318 and 61).
8). The relative positions of the character strings C and F and the specific characters B and E are indicated by arrows in the figure. Then, specific characters B and E are detected from the set position and area (steps 319 and 61).
9) it is normalized and sampled (steps 320, 62
0) and are identified (steps 321, 621).

If the sheet P is of the second type as shown in FIG. 23 and is conveyed upside down, the reverse character pattern in the character dictionary 30 is used (NO in step 317, N in step 617).
O). In this case, based on the positions of the character strings C and F, the positions and areas of the specific characters B and E are set as shown in FIG. 27 (steps 322 and 622). Then, specific characters B and E are detected from the set position and area (step
323,623), which are normalized and sampled (step 3
24,624) and are identified (steps 325,625).

Then, each character identification result is output to the comprehensive judgment section 31.
The type and transport state of the paper sheet P are determined by the overall determination unit 31.

That is, based on the content of the identified character and the position where the character is described, the type and the transport state (upright /
(Facing / upright / inverted).

As described above, necessary characters on the paper sheet P can be accurately captured irrespective of the surrounding image conditions, and the type and transport state of the paper sheet P can be captured accurately.

[3] The third embodiment will be described.

As shown in FIG. 28, an image pickup section 41 is provided at a position facing the conveyance path of the paper sheet P. Imaging unit 41
Has a function of capturing an image of a paper sheet with infrared light and visible light, respectively, and includes a filter (for example, a red filter) for imaging with visible light.

The two-dimensional analog image signal of the sheet P picked up by the infrared light is converted by the A / D converter 42 into a multi-valued digital image signal. This multi-valued digital image signal is converted into a density difference calculating section 44 which is a density difference image forming means.
Supplied to

The two-dimensional analog image signal of the sheet P picked up by visible light is converted into a multi-valued digital image signal by the A / D converter 43. The multi-valued digital image signal is supplied to the density difference calculation unit 44 and is also binarized by the binarization unit 49 and supplied to the character detection cutout unit 49.

The density difference calculation section 44 forms a density difference image corresponding to the difference between the density of an image picked up by infrared light and the density of an image picked up by visible light. The formed density difference image is binarized by a binarization circuit 45 and supplied to a character detection unit (connected image detection means) 46.

The character detecting section 46 detects a continuous connected image among the density difference images. The detected connected image is supplied to a character string area extraction unit (character string area detecting means) 47, where a character string area is detected from the connected image.

The character detection cutout section 49 detects a continuous connected image from the image corresponding to the character string area detected by the character string area extracting section 47 among the images picked up by the visible light, and Image detecting means), and detects a character from the detected connected image (character detecting means). The detection result is supplied to the character identification unit 51 via the normalization / sampling unit 50.

The character discriminating section 51 is a character string judging / selecting section 2
The character of the character string selected in 6 is identified by collating with the character pattern in the character dictionary 32. The character dictionary 52 stores a normal character pattern and a reverse character pattern that are upside down for each character. Then, the identification result of the character identification section 51 is supplied to the overall determination section 51.

Next, the operation of the above configuration will be described.

FIG. 29 is a flowchart for explaining the operation, and FIG. 30 shows an example of the conveyed sheets P.

In the sheet P shown in FIG. 30, an image A, a character B, and a character string C are described at a substantially left position on the front surface, and an image D, a character E, and a character string F are described at a substantially right position. I have. In particular, it is assumed that the image A and the character strings C and F are printed in green, and the others are printed in black.

Each imaging section 41 performs main scanning in a direction orthogonal to the transport direction of the paper sheet P. The sub-scan is performed by the main scanning moving along with the conveyance, and the surface of the sheet P is imaged by infrared light and visible light, respectively. Thereby, an infrared image and a visible image are captured (step 70).
1,702).

In imaging with visible light, as shown in FIG.
Images of all colors are raised. In the imaging using infrared light, as shown in FIG. 31, the green image A and the character strings C and F are erased.

The density βij of each density of the visible image is subtracted from the density αij of each pixel of the infrared image to form a density difference image (absolute value) corresponding to the density difference (step 703,
704). In this density difference image, only the green image A and the character strings C and F stand out as shown in FIG.

The density difference image is binarized (step 70).
5) A connected image (the coordinates of the start point and the end point) is detected from the density difference image by labeling, and the height and width, which are the character sizes of the connected image, are measured (step 706).

If the measured height and width are within the allowable range compared to the reference character parameters, the position of the connected image is extracted as a character string area (step 70).
7).

On the other hand, the visible image is binarized (step 70).
9) With respect to the visible image, information on an area that is applicable from the visible image in FIG. 30 is extracted from the extracted character string area. The extracted visible image is a green image A
And character strings C and F. Then, a continuous connected image is detected from the extracted visible image by labeling, and the height and width, which are the character sizes of the detected connected image, are measured (step 709).

From the detected connected images, characters whose character size and character pitch are within an allowable range as compared with a predetermined range, for example, a reference value, are extracted (step 710). That is, it is determined whether or not the measured character size and character pitch are within an allowable range as compared with the reference value. If the measured character size and the character pitch are within the allowable range, it is determined that the connected image is a character area. Becomes

The extracted characters are normalized and sampled (step 711), and are identified by collation with a character pattern in the character dictionary 52. In addition, as described in the second embodiment with reference to FIG. 18, the positions and areas of the specific characters B and E are set based on the positions of the character strings C and F (as shown in FIG. 24). . Then, specific characters B and E are detected from the set position and area from the visible image of FIG. 30, and are normalized, sampled, and identified (step 71).
2).

As described above, necessary characters (green portion) on the paper sheet P can be accurately captured irrespective of the surrounding image conditions, and accordingly, the type and transport state of the paper sheet P can be accurately captured. .

Here, the reason why the green image A and the character strings C and F disappear in the infrared image will be described.

Since black (carbon black) has a property of absorbing infrared rays, the level of the black portion is different from that of the base (for example, white), but the color ink has a property of transmitting infrared rays. Irradiated infrared light passes through the color ink and is reflected by the base, and the reflected light passes through the color ink again, so that it looks like the color of the base (white). For this reason, black remains but green disappears.

Regarding the advantage of using the infrared image, as described above, since the green portion does not appear in the infrared image, the green portion is obtained by obtaining the difference between the infrared image and the visible image. In the visible image, on the other hand, a green portion appears at a predetermined level in the visible image. Therefore, even if the green portion is obtained by changing the threshold between the visible images, the dynamic range cannot be large, and the green portion is not obtained. It is hard to detect.

Whether ultraviolet rays can be used instead of infrared rays is not applicable because ultraviolet rays also reflect colors.

[4] A fourth embodiment will be described.

Here, as shown in FIG. 33, the configuration is such that the binarization section 48 and the character detection cutout section 49 are removed. Other configurations are the same as those of the third embodiment.

As for the operation, as shown in the flowchart of FIG. 34, step 713 is provided instead of steps 709 and 710. Other operations are the same as in the third embodiment.

That is, a character image is directly extracted from a density difference image formed by an infrared image and a visible image (step 713). The extracted characters are normalized and sampled (step 711), and identified by collating with a character pattern in the character dictionary 52 (step 71).
2).

In the above-described second to fourth embodiments, the height and width are detected from the labeling image as a method of extracting the character string area. These embodiments are also performed in the first embodiment. A method of obtaining a character string from a projection pattern may be used. Necessary characters (green portion) can be accurately captured regardless of the surrounding image conditions, and thus the type and transport state of the paper sheet P can be accurately captured.

[0113]

As described above, according to the present invention, an image of a paper sheet is taken, a specific area is set for the taken image, and it is determined whether a character area exists in the set specific area. Along with the determination, a solid connected image is detected from the captured images, and if it is determined that a character region exists, a character string is detected from the detected connected images,
Since it is configured to identify the characters of this detected character string,
It is possible to provide a paper sheet reader that can accurately capture necessary characters on paper sheets regardless of the surrounding image conditions.

Further, according to the present invention, since the determination means for determining the type and the transport state of the sheet according to the result of the character identification is provided, it is possible to accurately capture the type and the transport state of the sheet. Can provide a paper sheet reading device capable of performing the above.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment.

FIG. 2 is a flowchart for explaining the operation of the first embodiment.

FIG. 3 is a diagram illustrating an example of a paper sheet P in the first embodiment.

FIG. 4 is a diagram illustrating an example of a paper sheet P in the first embodiment.

FIG. 5 is a diagram illustrating an example of a paper sheet P in the first embodiment.

FIG. 6 is a diagram illustrating an example of a paper sheet P in the first embodiment.

FIG. 7 is a diagram showing the formation of a projection pattern in the first embodiment.

FIG. 8 is a view showing the formation of a projection pattern in the first embodiment.

FIG. 9 is a diagram showing the formation of a projection pattern in the first embodiment.

FIG. 10 is a diagram showing the formation of a projection pattern in the first embodiment.

FIG. 11 is a diagram showing the formation of a projection pattern in the first embodiment.

FIG. 12 is a diagram showing the formation of a projection pattern in the first embodiment.

FIG. 13 is a diagram showing the formation of a projection pattern in the first embodiment.

FIG. 14 is a diagram showing the formation of a projection pattern in the first embodiment.

FIG. 15 is a diagram for explaining a character size and a character pitch in each embodiment.

FIG. 16 is a block diagram showing the configuration of the second embodiment.

FIG. 17 is a flowchart for explaining the operation of the second embodiment.

FIG. 18 is a flowchart following FIG. 17;

FIG. 19 is a flowchart following FIG. 17;

FIG. 20 is a diagram illustrating an example of a paper sheet P in the second embodiment.

FIG. 21 is a diagram showing an example of a paper sheet P in the second embodiment.

FIG. 22 is a diagram illustrating an example of a paper sheet P in the second embodiment.

FIG. 23 is a diagram showing an example of a paper sheet P in the second embodiment.

FIG. 24 is a diagram showing a relative positional relationship between a character string and a specific character in the second embodiment.

FIG. 25 is a diagram showing a relative positional relationship between a character string and a specific character in the second embodiment.

FIG. 26 is a diagram showing a relative positional relationship between a character string and a specific character in the second embodiment.

FIG. 27 is a diagram showing a relative positional relationship between a character string and a specific character in the second embodiment.

FIG. 28 is a block diagram showing the configuration of the third embodiment.

FIG. 29 is a flowchart for explaining the operation of the third embodiment.

FIG. 30 is a view showing a visible image in the third embodiment.

FIG. 31 is a diagram showing an infrared image in the third embodiment.

FIG. 32 is a diagram showing a density difference image in the third embodiment.

FIG. 33 is a block diagram showing the configuration of the fourth embodiment.

FIG. 34 is a flowchart for explaining the operation of the fourth embodiment;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Image pick-up part 4 ... Labeling coordinate extraction part 5 ... Character extraction part 6 ... Projection area setting part 7 ... Character area judgment part 8 ... Character string judgment part 10 ... General judgment part 11 ... Character identification part 12 ... Character dictionary

Continuing on the front page (72) Inventor Hisashi Takahashi 70, Yanagicho, Saiwai-ku, Kawasaki-shi, Kanagawa Prefecture, Japan Inside the Toshiba Yanagimachi Factory (72) Inventor Toshio Sato 70 Yanagimachi, Yuki-ku, Kawasaki-shi, Kanagawa Prefecture Inside Toshiba Yanagicho Plant Co., Ltd. (72) Inventor Chitetsu Nagatani 70 Yanagicho, Yuki-ku, Kawasaki City, Kanagawa Prefecture Toshiba Soshio Engineering Co., Ltd. Satoshi 70, Yanagicho, Saiwai-ku, Kawasaki City, Kanagawa Prefecture (72) Inventor Kei Takizawa Kei 70, Yanagicho, Kawasaki-shi, Kanagawa Prefecture Inside Toshiba Yanagicho Factory (72) Inventor Kazuyo Hashiya Kawasaki, Kanagawa Prefecture 70, Yanagicho, Sachi-ku, Toshiba Yanagicho Plant Co., Ltd. CA03 CA08 CA11 DA03 DA17 DC08 DC09 DC37 DC40 DC45

Claims (13)

[Claims] 1. An image pickup means for picking up an image of a paper sheet, an area setting means for setting a specific area for an image picked up by the image pickup means, and a character area in a specific area set by the area setting means Character area determining means for determining whether or not there is a connected image detecting means for detecting a continuous connected image among images captured by the image capturing means; and a positive determination result of the character area determining means. In this case, a character string detecting means for detecting a character string in the connected image detected by the connected image detecting means, and a character identifying means for identifying a character of the character string detected by the character string detecting means. A paper sheet reading device, characterized in that: 2. An image pickup means for picking up images on both front and back sides of a sheet, an area setting means for setting a specific area for the images on both front and back images picked up by the image pickup means, Character area determining means for determining whether or not a character area exists in each specific area; connecting image detecting means for detecting a continuous connected image among images captured by the image capturing means; When the determination result of the means is affirmative, a character string detecting means for detecting a character string in the connected image detected by the connected image detecting means, and a character of the character string detected by the character string detecting means is identified. A paper sheet reading device, comprising: character identification means. 3. The paper sheet reading device according to claim 1, wherein the character string detection unit detects the character string detection unit with the connected image detection unit when the determination result of the character region determination unit is affirmative. A character detecting means for detecting, as a character, a character which satisfies a predetermined condition among the connected images to be obtained, and a character for determining, as a character string, the presence of a character which satisfies a predetermined condition among the characters detected by the character detecting means A sheet reading device comprising: a column determining unit. 4. The paper sheet reading device according to claim 1, further comprising: a determination unit configured to determine a type, a conveyance state, and the like of the paper sheet according to an identification result of the character identification unit. A paper sheet reading device, characterized in that: 5. The paper sheet reading device according to claim 1, further comprising a character dictionary storing a normal character pattern and a reverse character pattern in an upside down direction for each character, Wherein the character is identified by collating with a character pattern in the character dictionary. 6. An image pickup means for picking up an image of a paper sheet, a divided area setting means for setting a plurality of divided areas for an image picked up by the image pickup means, A character string detecting means for detecting a character string from the divided area; a selecting means for selecting a character string meeting a specific condition among the character strings detected by the character string detecting means; a character selected by the selecting means A sheet identification device comprising: a character identification unit that identifies characters in a row; and a determination unit that determines a type, a conveyance state, and the like of the sheet in accordance with an identification result of the character identification unit. apparatus. 7. An image pickup means for picking up an image of a paper sheet, a divided area setting means for setting a plurality of divided areas for an image picked up by the image pickup means, A character string detecting means for detecting a character string from the divided area; a selecting means for selecting a character string that meets a specific condition from among the character strings detected by the character string detecting means; A character dictionary storing a character pattern and a reverse character pattern, first character identification means for identifying a character of a character string selected by the selection means by comparison with a character pattern in the character dictionary, and Character detecting means for detecting a specific character among the detected characters in accordance with the direction of the collated character pattern used by the first character identifying means; and identifying the specific character detected by the character detecting means. 2
A sheet identification device, comprising: a character identification unit; and a determination unit that determines a type, a conveyance state, and the like of the sheet according to identification results of the first and second character identification units. . 8. The paper sheet reading device according to claim 6, wherein the character string detecting means includes a character detecting means for detecting a character from each of the divided areas set by the divided area setting means. And a character string determining unit that determines, as a character string, the presence of a character that meets a predetermined condition among the characters detected by the character detecting unit. 9. An image pickup means for picking up an image of a sheet using infrared light and visible light, and a difference between a density of an image picked up by the infrared light and a density of an image picked up by the visible light. A density difference image forming means for forming a density difference image; a connected image detecting means for detecting a continuous connected image among the density difference images formed by the density difference image forming means; Character string region detecting means for detecting a character string region from the connected image, and among images captured by the visible light, an image corresponding to the character string region detected by the character string region detecting means, and Connected image detecting means for detecting a connected image, character detecting means for detecting a character from the connected image detected by the connected image detecting means, and a sentence identifying the character detected by the character detecting means A paper sheet reading device, comprising: character identification means. 10. An image pickup means for picking up an image of a sheet using infrared light and visible light, and a difference between a density of an image picked up by the infrared light and a density of an image picked up by the visible light. A density difference image forming means for forming a density difference image; a connected image detecting means for detecting a continuous connected image among the density difference images formed by the density difference image forming means; A character string area detecting means for detecting a character string area from the connected image, a character detecting means for detecting a character from a character string area detected by the character string area detecting means, and a character detected by the character detecting means. A sheet reading device comprising: a character identification unit for identifying; 11. The paper sheet reading device according to claim 9, further comprising a determination unit configured to determine a type, a conveyance state, and the like of the paper sheet according to an identification result of the character identification unit. A paper sheet reading device, characterized in that: 12. An image pickup means for picking up an image of a paper sheet, an area setting means for setting a specific area for the image picked up by the image pickup means, and a projection pattern from the specific area set by the area setting means. A projection pattern obtaining unit for obtaining the height of the projection pattern obtained by the projection pattern obtaining unit; and a measurement unit configured to determine whether the height measured by the measurement unit is within a predetermined range. Determining whether or not the projection pattern is a character area if the result of the determination is that the projection pattern is within a predetermined range. Based on the determination result by the determination means, A sheet reader for identifying characters in the projection pattern. 13. An image pickup means for picking up an image of a sheet using infrared light and visible light, and a difference between a density of an image picked up by the infrared light and a density of an image picked up by the visible light. Density difference image forming means for forming a density difference image; connecting image detecting means for detecting a connected image from the density difference image formed by the density difference image forming means; and the connection detected by the connected image detecting means Measuring means for measuring the height and width of the image; and determining whether the height and width measured by the measuring means are within a predetermined range. When it is determined that the projected pattern is a character area, the determination unit determines that the projected pattern is a character area, and, based on the determination result of the determination unit, identifies the characters in the connected image. Paper leaf Reader.