JP4805481B2 - Destination recognition device, sorter, destination recognition method and sorting method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention uses, for example, a destination recognition device and a destination recognition method for recognizing destination information from an image of a paper sheet such as a postal matter, and the destination recognition device or the destination recognition method. The present invention relates to a sorting machine and a sorting method for sorting paper sheets.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, a sorting machine has a destination recognition device that recognizes destination information described on a paper sheet such as a mail using a pattern recognition technique. The sorting machine having the destination recognition device sorts the paper sheets into predetermined sorting pockets based on the recognition result by the destination recognition device.
[0003]
A destination recognition device used in such a sorter detects an area in which destination information is described from a paper sheet image read from a paper sheet by a scanner or the like, and recognizes the destination information. ing. Furthermore, in the conventional destination recognition apparatus, a specific figure is detected before the destination information is detected in order to improve the detection accuracy of the destination information. A specific figure is a number information frame in which number information such as a zip code frame is written, a fee figure stamp such as a stamp, a fee meter and a postpayment stamp, a characteristic figure such as a surrounding frame, a dividing line, a photograph (specific figure) It is. By detecting these specific figures, in the destination information detection process, it is possible to predict the description position of the destination information or delete information unrelated to the destination information such as advertisements.
[0004]
For example, number information indicating a specific area such as a zip code is often written in the number information frame. For this reason, if the number information frame is detected, the number information can be easily detected. In addition, destination information is not described in the fee stamp, the advertisement described in the box, the advertisement described in the area indicated by the partition line, or the area where the photograph is described. In other words, the area in which the fee stamp or advertisement is written is an unnecessary part in the destination information recognition process, and if these areas are left as they are to detect the destination information, erroneous detection of the destination information will occur. It tends to happen. For this reason, the detection accuracy of the destination information can be improved by deleting the area detected as the fee stamp or the advertisement area. Here, information irrelevant to the destination information is called an advertisement, and an area where no destination information is described is called an advertisement area.
[0005]
Further, in the conventional destination recognition apparatus, an advertisement area is detected based on a surrounding frame, a partition line, or a photograph detected as a specific figure from a paper sheet image. For example, as shown in FIG. 24, an advertisement is often described in a box on a paper sheet. For this reason, when a surrounding frame is detected, the inside of the surrounding frame is determined as an advertisement area. In addition, as shown in FIG. 25, the partition lines on the paper sheets are often described as straight lines or dotted lines as the boundary between the destination information description area and the advertisement area. For this reason, when a partition line is detected, the area below the partition line with respect to the description direction of the destination information is determined as the advertisement area. In addition, as shown in FIG. 26, a photograph or an illustration may be printed on the surface on which the destination information is written on the paper sheet. Since such photos and illustrations (photo areas) are irrelevant to the destination information, when a photo area is detected, the area is determined as an advertisement. As described above, in the conventional destination recognition apparatus, an effort is made to improve the detection accuracy of the destination information by detecting various advertisement areas described on the paper sheet.
[0006]
However, in actual paper sheets, there are not only the above-mentioned types but also a great variety of advertisements. For this reason, in order to improve the recognition accuracy of destination information, it is desired to further increase the types of advertisements that can be detected.
[0007]
For example, in the conventional destination recognition device, the following cannot be detected as an advertisement.
(1) Advertisement that exists outside the box
As shown in FIG. 27, the advertisement is surrounded by a surrounding frame, but there is a paper sheet in which a part of the advertisement exists also outside the surrounding frame. In the conventional destination recognition device, the advertisement inside the frame can be removed, but the advertisement outside the box cannot be removed. For this reason, as shown in FIG. 28, there is a possibility that the character string of the destination is erroneously detected with the advertisement existing outside the surrounding frame as part of the destination information.
(2) Advertisement with customer barcode as boundary with destination information
In recent years, the number of paper sheets on which a barcode (customer barcode) in which destination information is coded is increasing. As shown in FIG. 29, a destination information description area, an advertisement description area, and a customer barcode are displayed. There are paper sheets that are separated. Even in the case of such paper sheets, the conventional destination recognition apparatus cannot detect an advertisement, and as shown in FIG. 30, an advertisement delimited by a customer barcode is used as a destination character line. May be detected by mistake.
[0008]
[Problems to be solved by the invention]
As described above, the present invention has been made in view of the above problems, and by improving the performance of detecting an area in which information unrelated to destination information is described from an image of a paper sheet, An object of the present invention is to provide a destination recognition device and a destination recognition method capable of improving information detection accuracy, and a sorting machine and a sorting method capable of improving processing capability using the destination recognition device or the destination recognition method. And
[0009]
Further, the present invention provides a destination recognition apparatus and a destination recognition method capable of improving the detection accuracy of destination information for an image of a paper sheet in which information irrelevant to the destination information exists outside the enclosing frame, It is an object of the present invention to provide a sorting machine and a sorting method capable of improving the processing capability using the destination recognition device or the destination recognition method.
[0010]
Further, the present invention provides a destination recognition device capable of improving the detection accuracy of destination information for a paper sheet image on which information irrelevant to the destination information is described using a customer barcode as a boundary with the destination information. It is another object of the present invention to provide a destination recognition method, a sorter and a sort method capable of improving processing capability using the destination recognition device or the destination recognition method.
[0011]
[Means for Solving the Problems]
The destination recognition apparatus according to the present invention is a destination recognition apparatus for recognizing destination information described on a paper sheet, a bar code detecting means for detecting a bar code indicating the destination from an image on the paper sheet, and the bar code If the barcode is detected by the detection means, If the position of the barcode in the longitudinal direction of the paper sheet is below the half position of the entire paper sheet, if the position of the barcode is below the predetermined position, the position of the barcode If there is a pixel distribution that seems to be destination information on the upper side, the position of the barcode is valid. A judging means for judging, and when the judging means judges that the position of the bar code is valid, Lower Deleting means for deleting the image of the area from the image on the paper sheet, and this deleting means Lower Recognizing means for recognizing destination information from the image of the paper sheet from which the image of the area is deleted.
[0012]
The sorting machine according to the present invention is a sorting machine that sorts paper sheets based on destination information described in the paper sheets, a transport unit that transports the paper sheets one by one, and a transport unit that transports the paper sheets one by one. A reading means for reading an image on a paper sheet to be read, a barcode detection means for detecting a barcode indicating a destination from an image read by the reading means, and a barcode is detected by the barcode detection means, If the position of the barcode in the longitudinal direction of the paper sheet is below the half position of the entire paper sheet, if the position of the barcode is below the predetermined position, the position of the barcode If there is a pixel distribution that seems to be destination information on the upper side, the position of the barcode is valid. A judging means for judging, and when the judging means judges that the position of the bar code is valid, Lower Deleting means for deleting the image of the area from the image on the paper sheet, and this deleting means Lower Recognizing means for recognizing destination information from the image of the paper sheet from which the image of the area has been deleted, and sorting means for sorting the paper sheets conveyed by the conveying means based on the destination information recognized by the recognizing means; It has.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a diagram showing a schematic configuration of a sorting apparatus according to the sorting machine of the present invention. As shown in FIG. 1, the sorting apparatus 1 includes a supply unit 11, a conveyance path (conveying unit) 12, a scanner (reading unit) 13, a destination recognition unit (destination recognition unit) 14, and a sorting pocket unit (sorting unit) 15. Etc.
[0016]
Destination information such as a zip code or address is written on the first surface of the supply unit 11 or the rear end is aligned so that the first surfaces of a plurality of printed paper sheets face the same direction. It is housed in the place. The supply unit 11 sequentially supplies the stored paper sheets to a predetermined take-out position. At the paper sheet take-out position, a take-out unit is provided for taking out the postal matter supplied by the supply unit 11 one by one along the main transport path 12 at a predetermined interval. Note that the paper sheets accommodated in the supply unit 11 are aligned so that the number information such as the zip code is positioned upward when the paper sheet is supplied to the take-out position, and the first surface faces the take-out unit. Has been placed.
[0017]
In addition to the paper sheets input via the supply unit 11 having only the text information (number information, location information, name, etc.) corresponding to the destination information on the first side, the first side There is a code in which a code (hereinafter referred to as a customer code) which is determined in correspondence with the destination information and is described in advance by the user of the paper sheet is described.
[0018]
The paper taken out from the supply unit 11 is conveyed through a conveyance path 12 having a conveyance belt that travels at a constant speed. On the transport path 12, there is provided a foreign matter / hardness detection unit for detecting the foreign matter and detecting the hardness of the paper sheet itself when the paper leaf contains foreign matter. On the downstream side of the foreign matter / hardness detection unit, an exclusion stacking unit is provided that excludes paper sheets that are determined to be incapable of mechanical processing by the foreign matter / hardness detection unit.
[0019]
A scanner 13 for reading an image of the first surface of the paper sheet on which the destination information is recorded is provided on the conveyance path 12 on the downstream side of the foreign matter / hardness detection unit. The scanner 13 is a scanning unit that optically reads and scans the first surface of the paper sheet. On the transport path 12 on the downstream side of the scanner 13, a positioning unit is provided for positioning the transport direction of the paper sheets. On the downstream side of the alignment unit, a barcode processing unit is provided that prints or reads a classification barcode on the first surface of the paper sheet as necessary.
[0020]
At the end of the conveyance path 12 on the downstream side of the barcode processing unit, a sorting pocket unit 15 for sorting and stacking paper sheets into predetermined sorting pockets (sorting destinations) is provided.
The division pocket portion 15 has a plurality of division pockets divided into a plurality of stages and a plurality of rows. For example, it is composed of 200 divided pockets 15a,. A plurality of switching gates are provided at the rear end of the transfer path 12, and a sorting gate (not shown) is provided above each sorting pocket. By selectively switching these sorting gates and switching gates, paper sheets are taken from the transport path 12 into each sorting pocket and accumulated.
Further, on the right side of the sorting apparatus 1 in the figure, a destination recognition unit 14 that performs destination information recognition processing and the like is provided together with a control unit that controls the operation of the sorting apparatus 1.
[0021]
FIG. 2 is a flowchart for explaining the paper sheet sorting process.
First, paper sheets to be processed are set in a predetermined position of the supply unit in a predetermined direction by an attendant. The sheets set in the supply unit 11 are supplied one by one onto the conveyance path 12 and conveyed to the scanner 13 (step S1). The scanner 13 inputs an image by reading and scanning paper sheets conveyed on the conveyance path 12 (step S2). The image input by the scanner 13 is sent to the destination recognition unit 14, and address recognition processing is performed using a pattern recognition technique (step S3). On the other hand, the paper sheets on the conveyance path 12 are classified into the respective classification pockets 15a,... Based on the reading result of the address by the destination recognition unit 14 (step S4).
[0022]
As described above, the sorting apparatus 1 sorts the paper sheet by recognizing the destination described on the paper sheet. The details of the entire sorting process are described in, for example, “Toshiba Review vol. 48, No. 7, p536-539 (1993)”.
[0023]
Next, destination information recognition processing by the destination recognition unit 14 according to the destination recognition apparatus of the present invention will be described.
FIG. 3 is a flowchart for explaining destination information recognition processing by the destination recognition unit 14. As shown in FIG. 3, the destination information recognition process is performed according to the following flow.
[0024]
First, the destination recognition unit 14 performs a specific graphic detection and an advertisement detection process based on the specific graphic (step S11). This specific graphic detection process is a process for increasing the accuracy of detection or recognition of destination information. In other words, in specific figure detection processing, characteristic figures such as postage code and other number information frames, postage stamps, charge stamps such as postage stamps, postage stamps, postage stamps, enclosing frames, divider lines, photographs, etc. (Specific figure) is detected. Based on the specific figure detected in this way, the destination recognition unit 14 performs an advertisement detection process. This advertisement detection process will be described in detail later. By detecting these specific figures, the destination recognition unit 14 may predict the description position of the destination information or delete (delete) an area in which an advertisement as information irrelevant to the destination information is described. it can. Here, information irrelevant to the destination information is called an advertisement, and an area in which the advertisement is described is called an advertisement area.
[0025]
Next, the destination recognition unit 14 performs destination information recognition processing (steps S12 to S16, recognition means). In the destination information recognition process, the destination recognition unit 14 first performs a process of detecting an area (destination area) in which the destination information is described (step S12). In this destination area detection process, an area where the destination area is likely to exist is detected while taking into account the position of the specific figure detected in the specific figure detection process.
[0026]
When the destination area is detected, the destination recognition unit 14 performs a destination character line detection process (step S13). In the destination character line detection process, the destination character line is detected from the destination area detected in the destination area detection process. When the destination character line is detected, the destination recognition unit 14 performs a character detection process (step S14). In this character detection process, individual characters are detected from the character line detected as the destination character line.
[0027]
When each character is detected, the destination recognition unit 14 performs character recognition processing (step S15). In this character recognition processing, character recognition processing is performed on each character detected in the character detection processing. When each character is recognized, the destination recognition unit 14 comprehensively recognizes the entire destination based on the recognition result of each character (step S16). In this destination recognition, the destination is determined by comparing the character recognition result obtained in the character recognition process with a destination dictionary such as an address dictionary (not shown).
[0028]
Next, the advertisement detection process will be described in detail.
FIG. 4 is a flowchart for explaining the advertisement detection process.
As shown in FIG. 4, first, the destination recognition unit 14 binarizes the entire sheet image read by the scanner 13 (step S21). The destination recognition unit 14 extracts connected components of black pixels from the binarized image (step S22). For example, when the scanner 13 reads an image as shown in FIG. 5, the destination recognition unit 14 extracts a black pixel connected component as shown in FIG. However, in FIG. 6, each black pixel connected component is displayed in a rectangle circumscribing it.
[0029]
When the black pixel connected components are extracted, the destination recognizing unit 14 detects various advertisement areas by examining each black pixel connected component (steps S23, S26, S29, S32, first and second detections). Means), the detected advertisement area is deleted (steps S25, S28, S31, S34, deletion means). First, the destination recognition unit 14 performs a photographic area detection process (step S23). This photographic region detection process will be described in detail later. When a photo area is detected by this photo area detection process (step S24, YES), the destination recognition unit 14 detects an advertisement area based on the detected photo area, That The advertisement area is erased from the paper sheet image (step S25).
[0030]
Next, the destination recognition unit 14 performs a partition line detection process (step S26). The partition line detection process will be described in detail later. When a partition line is detected by this partition line detection process (step S27), the destination recognition unit 14 detects an advertisement area based on the detected partition line with respect to the description direction of the destination information in the paper sheet image. The advertisement area is erased from the paper sheet image (step S28).
[0031]
Next, the destination recognition unit 14 performs a customer barcode detection process (step S29, barcode detection means). The customer barcode detection process will be described later in detail. When the customer barcode is detected by the customer barcode detection process (step S30), the destination recognition unit 14 detects an advertisement area based on the detected customer barcode, and erases the advertisement area from the paper sheet image. (Step S31).
[0032]
Next, the destination recognition unit 14 performs a surrounding frame detection process (step S32). The enclosing frame detection process will be described in detail later. When a surrounding frame is detected by this surrounding frame detection process (step S33), the destination recognition unit 14 detects an advertising area based on the detected surrounding frame, and erases the advertising area from the image of the paper sheet ( Step S34). The destination recognition unit 14 detects a black pixel connected component existing in a predetermined area near the surrounding frame as an advertisement area.
[0033]
By the processing as described above, the advertisement area is deleted based on the photograph, the partition line, the customer barcode, and the surrounding frame. Thereby, the destination information detection process is performed on the image in which the advertisement area is erased, and the destination area can be detected with high accuracy.
[0034]
Next, the photographic area detection process will be described in detail.
FIG. 7 is a flowchart for explaining the flow of the photo detection process.
First, the destination recognition unit 14 binarizes the image read by the scanner 13 and extracts the number of black pixel connected components. Here, the destination recognizing unit 14 defines the number of extracted black pixel connected components as Np, defines the variable Ip, and sets the initial value of the variable Ip to 0 (step S41). The destination recognition unit 14 compares the variable Ip with the number Na of black pixel connected components and determines whether Ip> Np (step S42). If Ip> Np is not determined by this determination, the destination recognizing unit 14 sets the coordinates of a circumscribed rectangle surrounding the black pixel connected component to be subjected to the photo detection process (step S43). When the coordinates of the circumscribed rectangle surrounding the black pixel connected component are set, the destination recognition unit 14 calculates the size of the black pixel connected component from the set coordinates (step S44).
[0035]
For example, in the example shown in FIG. 8, if the short direction of the paper sheet is the X direction and the long direction is the Y direction, the coordinates of the circumscribed rectangle surrounding the black pixel connected component are (Xp1, Xp2, Yp1, Yp2). ing. Here, assuming that the length in the X direction of the circumscribed rectangle surrounding the black pixel connected component is Xp and the length in the Y direction is Yp, the black pixel connected components shown in FIG. 8 are Xp = Xp2-Xp1, Yp = Yp2- Yp1.
[0036]
When the size of the black pixel connected component is calculated, the destination recognizing unit 14 compares the predetermined threshold value with the size of the black pixel connected component, and the size of the black pixel connected component is appropriate as a photographic region candidate. Whether or not (step S45). That is, when it is determined that the size of the black pixel connected component is appropriate (step S45, YES), the destination recognition unit 14 sets the black pixel connected component as a photographic region candidate. If the size of the black pixel connected component is not appropriate, the destination recognizing unit 14 does not determine the black pixel connected component as a photographic area candidate (NO in step S45), and sets the variable Ip = Ip + 1 to the next black. The processing shifts to the pixel connected component (step S49).
[0037]
Here, the threshold values for determining a photograph area candidate are XMINp (minimum value in the X direction), XMAXp (maximum value in the X direction), YMINp (minimum value in the Y direction), and YMAXp (maximum value in the Y direction). Suppose that it is set. Then, the destination recognition unit 14 determines whether or not the size of the black pixel connected component is appropriate by determining whether or not the black pixel connected component has a size within the range of the threshold. For example, the black pixel connected component shown in FIG.
XMINp <Xp <XMAXp, YMINp <Yp <YMAXp
If these two expressions are satisfied, the black pixel connected component is determined as a photographic region candidate.
[0038]
The destination recognition unit 14 calculates the density of black pixels for the black pixel connected component determined to be a photographic region candidate (step S46). The black pixel density is calculated based on the density value of each pixel in the black pixel connected component area read by the scanner 13. For example, the destination recognition unit 14 compares the density value d of each pixel included in the area with a predetermined threshold value Tp, and sets the number of pixels satisfying d> Tp as Cp. This Cp gives the number of black pixels equal to or higher than a predetermined density value included in the photographic region candidate. Therefore, the destination recognition unit 14 determines whether or not the ratio of black pixels that are equal to or higher than a predetermined density value included in the photographic area candidate is larger than a predetermined value (step S47).
[0039]
The destination recognizing unit 14 determines that the ratio of black pixels having a density value equal to or higher than the threshold value Tp included in the photo area candidate is larger than a predetermined threshold value Sp (YES in step S47), that is, Cp / (Xp × Yp). )> When Sp is satisfied, the photographic area candidate is determined as a photographic area. If it is determined that the black pixel connected component is a photographic region, the destination recognition unit 14 detects the black pixel connected component as a photographic region (step S48), and processes the next black pixel connected component as a variable Ip = Ip + 1. (Step S49).
[0040]
In addition, when the ratio of black pixels having a density value equal to or higher than the threshold value Tp included in the photo area candidate is not greater than the predetermined threshold value Sp (NO in step S47), the destination recognition unit 14 determines that the photo area candidate is a photo area. Is determined, and the process proceeds to the process for the next black pixel connected component with the variable Ip = Ip + 1 (step S49).
[0041]
Next, an advertisement area detection process using a partition line will be described in detail.
FIG. 9 is a flowchart for explaining the flow of the partition line detection process.
Here, since the dividing line is often printed in the short direction of the paper sheet, in order to detect the dividing line, a straight line extending in the X direction is detected. Here, the image of the paper sheet shown in FIG. 10 will be described as an example.
[0042]
First, the destination recognition unit 14 binarizes the image read by the scanner 13 and extracts the number of black pixel connected components. Here, the destination recognition unit 14 defines Ns as the number of extracted black pixel connected components, defines the variable Is, and sets the initial value of the variable Is to 0 (step S51). The destination recognition unit 14 compares the variable Is with the number Ns of black pixel connected components, and determines whether Is> Ns (step S52). If Is> Ns is not determined by this determination, the destination recognizing unit 14 sets the coordinates of a circumscribed rectangle surrounding the black pixel connected component (step S53). When the coordinates of the circumscribed rectangle surrounding the black pixel connected component are set, the destination recognition unit 14 calculates the size of the black pixel connected component from the set coordinates (step S54).
[0043]
When the size of the black pixel connected component is calculated, the destination recognition unit 14 compares the predetermined threshold value with the size of the black pixel connected component, and determines whether the size of the black pixel connected component is appropriate as a partition line. It is determined whether or not (step S55). When the destination recognition unit 14 determines that the size of the black pixel connected component is appropriate as the partition line, the destination recognition unit 14 sets the black pixel connected component as the partition line region candidate.
[0044]
For example, in the example shown in FIG. 10, the transverse direction of the rectangle surrounding the black pixel connected component is (Xs1, Xs2, Ys1, Ys2), where the short direction of the paper sheet is the X direction and the long direction is the Y direction. It shall be. Here, if the length in the X direction of the circumscribed rectangle surrounding the black pixel connected component is Xs and the length in the Y direction is Ys, the black pixel connected component is Xs = Xs2-Xs1, Ys = Ys2-Ys1. . These Xs and Ys are compared with predetermined threshold values XMINs (minimum value in the X direction), XMAXs (maximum value in the X direction), YMINs (minimum value in the Y direction), and YMAXs (maximum value in the Y direction). Thus, it is determined whether or not the black pixel connected component is a straight line long in the X direction. For example, it is determined whether the threshold value is small in the Y direction and large in the X direction so that it can be determined whether the partition line is written in the short direction of the paper sheet. It is set to judge.
[0045]
That is, the destination recognition unit 14 performs the above-described Xs and Ys
XMINs <Xs <XMAXs, YMINs <Ys <YMAXs
If the equation (2) is satisfied, the black pixel connected component is determined as a partition line candidate.
[0046]
Next, the destination recognizing unit 14 determines whether or not the black pixel connected component determined to be a partition line candidate exists at an appropriate position on the paper sheet (step S56). Here, the appropriate position of the partition line candidate is a reference for determining whether or not the lower side of the partition line is an advertisement area that does not include destination information. Here, “lower side” refers to a region having a Y coordinate larger than the maximum Y coordinate (Ys2) of the partition line candidate.
[0047]
For example, when the position of the partition line candidate in the Y direction is below half of the entire sheet, the area positioned below the partition line candidate may be determined as the advertisement area. In addition, when the position of the partition line candidate is below the predetermined position, the lower side of the partition line may be determined as the advertisement area. Further, when there is a pixel distribution that seems to be destination information above the position of the candidate for the partition line, the lower side of the partition line may be determined as the advertisement area.
[0048]
If it is determined by the above determination that the position of the partition line candidate is appropriate (YES in step S56), the destination recognition unit 14 detects the partition line candidate as a partition line (step S57), and sets the variable Is = Is + 1 as the next. The process proceeds to the process for the black pixel connected component (step S58).
[0049]
If it is determined by the above determination that the position of the partition line candidate is not valid (step S56, NO), the destination recognition unit 14 does not detect the partition line candidate as a partition line and sets the variable Is = Is + 1 as The process proceeds to the process for the black pixel connected component (step S58).
[0050]
Next, an advertisement area detection process using a surrounding frame will be described in detail.
FIG. 11 is a flowchart for explaining the flow of the surrounding frame detection process.
First, the destination recognition unit 14 binarizes the image read by the scanner 13 and extracts the number of black pixel connected components. Here, the destination recognizing unit 14 defines the number of extracted black pixel connected components as Nb, defines the variable Ib, and sets the initial value of the variable Ib to 0 (step S61). The destination recognition unit 14 compares the variable Ib with the number Nb of black pixel connected components, and determines whether Ib> Nb is satisfied (step S62). If it is determined that Ib> Nb is not satisfied, the destination recognition unit 14 sets the coordinates of a circumscribed rectangle surrounding the black pixel connected component that is the target of the surrounding frame detection process (step S63). When the coordinates of the circumscribed rectangle surrounding the black pixel connected component are set, the destination recognition unit 14 calculates the size of the black pixel connected component from the set coordinates (step S64).
[0051]
When the size of the black pixel connected component is calculated, the destination recognition unit 14 compares the predetermined threshold value with the size of the black pixel connected component, and the size of the black pixel connected component is appropriate as the enclosing frame. Whether or not (step S65). When the destination recognizing unit 14 determines that the size of the black pixel connected component is appropriate as the surrounding frame, the destination recognizing unit 14 sets the black pixel connected component as the surrounding frame region candidate.
[0052]
For example, in the example shown in FIG. 11, the lateral direction of the rectangle surrounding the black pixel connected component is (Xb1, Xb2, Yb1, Yb2), where the short direction of the paper sheet is the X direction, the longitudinal direction is the Y direction. It shall be. Here, if the length in the X direction of the circumscribed rectangle surrounding the black pixel connected component is Xb and the length in the Y direction is Yb, the black pixel connected component is Xb = Xb2-Xb1, Yb = Yb2-Yb1. . These Xb and Yb are compared with predetermined threshold values XMINb (minimum value in the X direction), XMAXb (maximum value in the X direction), YMINb (minimum value in the Y direction), and YMAXb (maximum value in the Y direction). Thus, it is determined whether or not the black pixel connected component is a region that seems to be a surrounding frame.
[0053]
That is, the destination recognition unit 14 performs the above-described Xb and Yb.
XMINb <Xb <XMAXb, YMINb <Yb <YMAXb
If the equation (2) is satisfied, the black pixel connected component is set as a surrounding frame region candidate.
[0054]
Next, the destination recognizing unit 14 determines whether or not the black pixel connected component determined to be a surrounding frame region candidate exists at an appropriate position on the paper sheet (step S66). Here, the appropriate position of the surrounding frame region candidate is a reference for determining whether or not the advertising region does not include the destination information in the surrounding frame or below the surrounding frame. Here, “lower side” indicates an area having a larger Y coordinate than the minimum value (Yb1) of the Y coordinate of the surrounding frame area candidate.
[0055]
For example, when the position in the Y direction of the surrounding frame area candidate exists below half of the entire paper sheet, the area positioned below the surrounding frame area candidate may be determined as the advertisement area. Alternatively, when the position of the surrounding frame area candidate is below the predetermined position, the lower side of the surrounding frame may be determined as the advertisement area. Further, when there is a pixel distribution that seems to be destination information above the position of the surrounding frame region candidate, the lower side of the surrounding frame may be determined as the advertising region.
[0056]
If it is determined by the above determination that the position of the surrounding frame region candidate is appropriate (step S66, YES), the destination recognizing unit 14 calculates a projection value of the surrounding frame region candidate in the X direction and the Y direction (step). S67). For example, in the example shown in FIG. 12, a projection value having a peak at the end point of the surrounding frame region candidate is obtained. When the region is a surrounding frame, the projection has peaks at end points in the X and Y directions of the surrounding frame. Therefore, the destination recognizing unit 14 determines whether or not the frame is a frame by determining whether or not there is a projection peak at the end points in the X direction and the Y direction of the frame region candidate (step S68).
[0057]
In addition, when the peak exists only at the end point in the X direction, or when the peak exists only at the end point in the Y direction, the surrounding frame candidate region may be determined as the surrounding frame. This is because there may be a line indicating a surrounding frame only in the X direction or the Y direction.
[0058]
If there is a projection peak at the end point of the surrounding frame region candidate, the destination recognition unit 14 determines that the surrounding frame region candidate is a surrounding frame (step S68), and detects the region as a surrounding frame (step S69). Then, the process proceeds to the process for the next black pixel connected component with the variable Is = Is + 1 (step S70). If there is no projection peak at the end point of the surrounding frame area candidate as a result of the determination (step S68, NO), the destination recognizing unit 14 does not detect the surrounding frame area candidate as the surrounding frame, and the variable Is = Is + 1 Then, the process proceeds to the process for the next black pixel connected component (step S70).
[0059]
When the surrounding frame is detected, the destination recognition unit 14 detects an advertising area around the surrounding frame. FIG. 13 is a flowchart for explaining the detection process of the advertisement area related to the surrounding frame. FIG. 14 shows an example of a paper sheet image in which a date (advertisement) irrelevant to the destination information is described in the upper part of the surrounding frame.
[0060]
That is, when a surrounding frame is detected by the surrounding frame detection process, the destination recognition unit 14 detects an advertisement area based on the detected surrounding frame. First, the destination recognizing unit 14 determines an area inside the enclosing frame as an advertising area, and detects the inside of the enclosing frame as an advertising area (first area) (step S71, first detecting means).
[0061]
Further, the destination recognition unit 14 determines whether or not there is a character string as an advertisement unrelated to the destination information around the surrounding frame (steps S72 to S75, second detection means). That is, there may be a description that is not related to the destination information such as the date around the frame. For this reason, the destination recognizing unit 14 determines that a description unrelated to the destination information existing around the surrounding frame is also an advertisement area. For example, in the example shown in FIG. 14, a date unrelated to the destination information is described in the upper part of the surrounding frame. As shown in FIG. 14, if there is a description unrelated to the destination information at the top of the enclosing frame, the destination recognition unit 14 often erroneously detects the destination information. For this reason, the destination recognizing unit 14 sets a predetermined inspection area around the surrounding frame (step S72), and determines whether or not an advertisement exists.
[0062]
Here, as shown in FIG. 14, an inspection region having a predetermined height (length in the Y direction) is set at the upper part of the surrounding frame. The height of the inspection area is a predetermined length Rb. The inspection area is not limited to the upper part of the surrounding frame. For example, the right side of the surrounding frame, the left side of the surrounding frame, or the lower part of the surrounding frame may be set as the inspection area. When detecting an advertisement on the right side or the left side of the surrounding frame, an area having a predetermined length from the end portion in the X direction of the surrounding frame may be set as the inspection area.
[0063]
When an inspection area having a predetermined height is set in the upper part of the detected surrounding frame, the destination recognition unit 14 calculates the number Nk of black pixel connected components in the inspection area (step S73). When the number Nk of black pixel connected components in the inspection area is calculated, the destination recognition unit 14 determines whether the number Nk of black pixel connected components in the inspection area is equal to or greater than a preset constant Cb. It is determined whether or not (step S74).
[0064]
That is, if an advertisement character string such as a date exists in the inspection area, the black pixel connected component should be detected by the number of advertisement characters. On the other hand, when a part of destination information that is not an advertisement exists in the inspection area, a connected component of about 1 to 3 characters is only detected. Therefore, the destination recognition unit 14 can determine whether or not an advertisement exists in the inspection area by setting the constant Cb in advance.
[0065]
When the number Nk of the black pixel connected components is equal to or larger than the constant Cb (step S74, YES), the destination recognition unit 14 determines that there is an advertisement in the inspection area, and determines the inspection area as the advertisement area (second (Step S75). Thereby, as shown in FIG. 15, a surrounding frame and the said detection area are detected as an advertisement area. If the number Nk of black pixel connected components is not equal to or greater than the constant Cb (step S74, NO), the destination recognition unit 14 determines that there is no advertisement in the inspection area and does not detect the inspection area as the advertisement area. To.
[0066]
As described above, a surrounding frame is detected from the image of the paper sheet, a predetermined inspection area is set around the detected surrounding frame, it is determined whether an advertisement exists in the inspection area, and the advertisement is When it is determined that it exists, the inspection area is detected as an advertisement area. Accordingly, it is possible to detect an advertisement that exists around the surrounding frame together with the surrounding frame, and it is possible to improve the accuracy of the destination information detection process.
[0067]
In addition, when the number of connected components in the inspection area is equal to or greater than a predetermined number, the inspection area is detected as an advertisement area. Therefore, even if only one part of destination information exists in the inspection area, inspection is performed. An area is not detected as an advertisement area, and a highly accurate destination information detection process can be realized.
[0068]
In the above steps S72 to S75, the advertisement area detection processing from the periphery of the surrounding frame has been described. Similarly to the above, an advertisement area may be detected. That is, when one advertisement area is detected, an advertisement existing around the detected advertisement area can also be detected as the advertisement area by determining the presence / absence of an advertisement around the advertisement area.
[0069]
Next, an advertisement area detection process using a customer barcode (hereinafter simply referred to as a barcode) will be described in detail.
FIG. 16 is a flowchart for explaining the flow of the barcode detection process. Here, a case where a barcode printed along the X direction of a paper sheet is detected will be described. bar In the code detection process, first, each bar candidate constituting the bar code is detected, and among the detected bar candidates, adjacent bar candidates are integrated to obtain a bar code.
[0070]
That is, the destination recognition unit 14 binarizes the image read by the scanner 13 and extracts the number of black pixel connected components. Here, the destination recognition unit 14 defines the number of extracted black pixel connected components as Nc, defines the variable Ic, and sets the initial value of the variable Ic to 0 (step S81). The destination recognition unit 14 compares the variable Ic with the number Nc of black pixel connected components, and determines whether Ic> Nc (step S82). If Ic> Nc is not determined by this determination, the destination recognizing unit 14 sets the coordinates of the circumscribed rectangle surrounding the black pixel connected component to be subjected to the bar candidate detection process (step S83). When the coordinates of a circumscribed rectangle surrounding the black pixel connected component are set, the destination recognition unit 14 calculates the size of the black pixel connected component from the set coordinates (step S84).
[0071]
When the size of the black pixel connected component is calculated, the destination recognizing unit 14 compares a predetermined threshold value with the size of the black pixel connected component, and the bar candidate whose size is the black pixel connected component constitutes a barcode. It is judged whether it is appropriate as (step S85). If the destination recognition unit 14 determines that the size of the black pixel connected component is appropriate as a bar candidate, the destination recognition unit 14 sets the black pixel connected component as a bar candidate (step S86).
[0072]
For example, in the example shown in FIG. 17, the lateral direction of the rectangle surrounding the black pixel connected component is (Xc1, Xc2, Yc1, Yc2), where the short direction of the paper sheet is the X direction, the longitudinal direction is the Y direction. It shall be. Here, if the length in the X direction of the circumscribed rectangle surrounding the black pixel connected component is Xc and the length in the Y direction is Yc, the black pixel connected component is Xc = Xc2-Xc1, Yc = Yc2-Yc1. .
[0073]
These Xc and Yc are compared with predetermined threshold values XMINc (minimum value in the X direction), XMAXc (maximum value in the X direction), YMINc (minimum value in the Y direction), and YMAXc (maximum value in the Y direction). Thus, it is determined whether or not the black pixel connected component is a bar candidate. Since the bar candidates constituting the customer barcode are described in a predetermined size, the threshold value is set based on the size to be detected as a bar candidate. That is, the destination recognition unit 14 performs the above-described Xc and Yc.
XMINc <Xc <XMAXc, YMINc <Yc <YMAXc
If the equation (2) is satisfied, the black pixel connected component is set as a bar candidate.
[0074]
When determining that the black pixel connected component is a bar candidate, the destination recognition unit 14 detects the black pixel connected component as a bar candidate (step S86), and proceeds to processing for the next black pixel connected component as a variable Is = Is + 1. (Step S87). If it is determined that the black pixel connected component is not appropriate as a bar candidate (NO in step S85), the destination recognition unit 14 does not detect the black pixel connected component as a bar candidate, and the variable Is = The process proceeds to the process for the next black pixel connected component as Is + 1 (step S87). By performing the processing of steps S82 to S87 for all the black pixel connected components, the destination recognition unit 14 detects all the bar candidates on the image of the paper sheet.
[0075]
When it is determined in step S82 that Ic = Nc, that is, when the bar candidate detection processing for all the black pixel connected components is completed, the destination recognition unit 14 determines whether the adjacent bar candidates are among the detected bar candidates. Are integrated (step S89), and a barcode is detected (step S90).
[0076]
For example, FIG. 17 shows a case where it is determined whether or not the bar candidate A and the bar candidate B are integrated. For the bar candidate integration determination, an X-direction margin Xcm and a Y-direction margin Ycm are used as predetermined constants. That is, an integrated area having a length that is Ycm larger than the top of the bar candidate A and a length that is larger by Xcm on the left and right is set, and the bar candidate B included therein is integrated. In this way, barcodes are detected by integrating bar candidates one after another.
[0077]
When the barcode is detected, the destination recognition unit 14 performs an advertisement area detection process based on the detected barcode. FIG. 18 is a flowchart for explaining an advertisement area detection process based on a barcode.
In other words, the destination recognition unit 14 determines the position coordinates of the detected barcode (step S91). When the position coordinates of the barcode are determined, the destination recognition unit 14 determines that the detected barcode is a predetermined condition. The It is determined whether or not the bar code position is valid depending on whether or not it is satisfied (step S92). Here, the predetermined condition for determining whether or not the barcode position is valid is a criterion for determining whether or not the lower side of the barcode is an advertisement area that does not include destination information. Here, “lower side” indicates an area having a Y coordinate larger than the Y coordinate of the barcode position.
[0078]
For example, the appropriate position of the barcode on the paper sheet is lower than the barcode when the barcode position is below the half position of the entire paper sheet in the Y direction of the paper sheet. The position area may be determined to be the advertisement area, or when the barcode position is below the predetermined position, the lower side of the barcode may be determined to be the advertisement area. Further, when a pixel distribution that is likely to be destination information exists above the barcode position, the lower side of the barcode may be determined as the advertisement area.
[0079]
If it is determined by the above determination that the position of the barcode is valid (step S92, YES), the destination recognition unit 14 detects the lower area of the barcode as an advertisement area based on the barcode (step S93). Accordingly, as shown in FIG. 19, when an advertisement is written below the barcode, the area below the barcode can be detected as the advertisement area.
[0080]
If it is determined by the above determination that the position of the barcode is not valid (step S92, NO), the destination recognition unit 14 does not detect the area below the barcode as the advertisement area based on the barcode. This is because even if a barcode is detected, if the position of the barcode is not valid, the destination information may be described below the barcode.
[0081]
As described above, the barcode is detected from the image of the paper sheet, and when the detected barcode is present at the appropriate position, the lower area of the barcode is detected as the advertisement area. is there. Thereby, when the advertisement is described with the barcode as a boundary, the advertisement below the barcode can be deleted, and the accuracy in the destination information detection process can be improved.
[0082]
Further, the validity of the barcode position is determined by determining whether destination information is written below the barcode. As a result, when the destination information is described below the barcode, accurate destination information detection processing can be performed without detecting the area below the barcode as the advertisement area.
[0083]
Next, a flow of processing for detecting an advertisement area from an original image and detecting an area (destination area) in which destination information is described will be described. Here, in the destination area detection process, as described above, the destination area is detected using the advertisement area detected based on the photograph, the dividing line, the surrounding frame, and the customer barcode.
[0084]
20 to 23 are diagrams for explaining a process from reading an image on a paper sheet by the scanner 13 to detecting a destination area. FIG. 20 shows an example of an image (original image) read by the scanner 13. FIG. 21 shows an advertisement area detected from the original image shown in FIG. 22 shows an image obtained by deleting the advertisement area shown in FIG. 21 from the original image shown in FIG. FIG. 23 shows an area (destination area) in which destination information detected from an image obtained by deleting the advertisement area shown in FIG. 22 is described.
[0085]
First, in the sorting machine, an image of a paper sheet is read by the scanner 13. The destination recognition unit 14 performs the advertisement detection process as described above on the image read by the scanner 13. Next, the destination recognition unit 14 that has detected the advertisement area erases only the advertisement area from the image read by the scanner 13.
[0086]
When the advertisement area is deleted in this way, the destination recognition unit 14 detects a character string on the image by integrating adjacent black pixel connected components, and further detects a destination area candidate by integrating the character strings. . At this time, when there are a plurality of destination area candidates, the destination recognition unit 14 determines a destination area candidate close to the center on the paper sheet as the destination area. That is, if the destination area candidate is detected from an image in which descriptions other than the destination information are erased as much as possible, the detection accuracy of the destination area is improved.
[0087]
For example, when an image as illustrated in FIG. 20 is read by the scanner 13, the destination recognition unit 14 includes, as illustrated in FIG. 21, an area including an advertisement described in a box and an upper part of the box as an advertisement area. Is detected as an advertising area. When the destination recognition unit 14 deletes the advertisement area shown in FIG. 21, an image as shown in FIG. 22 is obtained. Furthermore, when the destination recognition unit 14 detects a destination area from the image shown in FIG. 22, a destination area as shown in FIG. 23 is detected.
[0088]
【The invention's effect】
As described above in detail, according to the present invention, it is possible to improve the detection accuracy of the destination information by improving the performance of detecting an area in which information unrelated to the destination information is described from the image of the paper sheet. A destination recognition apparatus and a destination recognition method that can be performed, and a sorting machine and a sorting method that can improve the processing capability by using the destination recognition apparatus or the destination recognition method can be provided.
[0089]
In addition, according to the present invention, a destination recognition device and destination recognition that can improve the detection accuracy of destination information for an image of a paper sheet in which information irrelevant to the destination information is also present outside the enclosing frame. It is possible to provide a sorting apparatus and a sorting method capable of improving the processing capability by using the method, the destination recognition apparatus or the destination recognition method.
[0090]
Further, according to the present invention, destination recognition that can improve the detection accuracy of destination information for an image of a paper sheet in which information irrelevant to the destination information exists with a customer barcode as a boundary with the destination information. It is possible to provide an apparatus and a destination recognition method, and a sorting machine and a sorting method capable of improving the processing capability using the destination recognition apparatus or the destination recognition method.
[Brief description of the drawings]
FIG. 1 is a diagram schematically showing a configuration of a sorting machine according to an embodiment of the present invention.
FIG. 2 is a flowchart for explaining a flow showing a processing flow in the sorting machine;
FIG. 3 is a flowchart for explaining a flow of destination recognition processing by a destination recognition unit;
FIG. 4 is a flowchart for explaining the overall flow of advertisement detection processing;
FIG. 5 is a diagram illustrating an example of a paper sheet image read by a scanner.
6 is a diagram showing an example of a black pixel connected component extracted from the paper sheet image of FIG. 5;
FIG. 7 is a flowchart for explaining the flow of a photographic region detection process.
FIG. 8 is a diagram for explaining coordinates of a black pixel connected component;
FIG. 9 is a flowchart for explaining the flow of a partition line detection process;
FIG. 10 is a diagram showing an example of an advertisement area detected by a partition line.
FIG. 11 is a flowchart for explaining the flow of a surrounding frame detection process;
FIG. 12 is a diagram showing an example of a projection value of an area detected as a surrounding frame.
FIG. 13 is a flowchart for explaining advertisement area detection processing based on an area detected as a surrounding frame;
FIG. 14 is a diagram showing an example of an image in which a character string as an advertisement exists in the upper part of a surrounding frame.
FIG. 15 is a diagram showing an advertisement area detected from the image shown in FIG. 14;
FIG. 16 is a flowchart for explaining the flow of bar code detection;
FIG. 17 is a diagram showing an example of bar candidates constituting a barcode.
FIG. 18 is a flowchart for explaining an advertising area detection process based on a detected barcode.
FIG. 19 is a diagram showing an example of an advertisement area detected based on a barcode.
FIG. 20 is a diagram illustrating an example of an image read by a scanner.
FIG. 21 is a view showing an advertisement area detected from the image of FIG. 20;
22 is a view showing an image obtained by deleting the advertisement area shown in FIG. 21 from the image of FIG.
23 is a diagram showing an example in which a destination area is detected from the image of FIG.
FIG. 24 is a diagram illustrating an example of an image of a paper sheet having a surrounding frame in which an advertisement is described.
FIG. 25 is a diagram illustrating an example of an image of a paper sheet in which an advertisement is described below the partition line.
FIG. 26 is a diagram illustrating an example of an image of a paper sheet on which a photograph is described.
FIG. 27 is a diagram illustrating an example of an image of a paper sheet in which a character string as an advertisement exists around a surrounding frame.
FIG. 28 is a diagram for explaining an example in which character line detection of destination information is erroneously detected by a character string as an advertisement around the frame shown in FIG. 27;
FIG. 29 is a diagram showing an example of an image of a paper sheet in which an advertisement is written below the customer barcode.
30 is a diagram for explaining an example in which character line detection of destination information is erroneously detected in an image of a paper sheet in which an advertisement is written below the customer barcode shown in FIG. 29;
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 11 ... Supply part, 12 ... Conveyance path (conveyance means), 13 ... Scanner (reading means), 14 ... Destination recognition part (destination recognition apparatus), 15 ... Sorting pocket part (sorting means)

Claims (4)

In a destination recognition device that recognizes destination information described on a paper sheet,
Barcode detecting means for detecting a barcode indicating a destination from an image on a paper sheet;
When the barcode is detected by the barcode detection means, the barcode position is a predetermined position when the barcode position is below the half position of the entire sheet in the longitudinal direction of the sheet. A determination means for determining that the position of the barcode is valid when a pixel distribution that is likely to be destination information exists above the position of the barcode ,
If the determination means determines that the position of the barcode is valid, a deletion means for deleting the image of the area below the barcode from the image on the paper sheet,
Recognizing means for recognizing destination information from the image of the paper sheet from which the image of the lower area of the barcode is deleted by the deleting means;
A destination recognition device comprising: In a sorting machine that sorts paper sheets based on the destination information written on the paper sheets,
A conveying means for conveying the paper sheets one by one;
Reading means for reading an image on a paper sheet conveyed by the conveying means;
Barcode detecting means for detecting a barcode indicating a destination from an image read by the reading means;
When the barcode is detected by the barcode detection means, the barcode position is a predetermined position when the barcode position is below the half position of the entire sheet in the longitudinal direction of the sheet. A determination means for determining that the position of the barcode is valid when a pixel distribution that is likely to be destination information exists above the position of the barcode ,
If the determination means determines that the position of the barcode is valid, a deletion means for deleting the image of the area below the barcode from the image on the paper sheet,
Recognizing means for recognizing destination information from the image of the paper sheet from which the image of the lower area of the barcode is deleted by the deleting means;
Sorting means for sorting paper sheets conveyed by the conveying means based on destination information recognized by the recognizing means;
A sorting machine characterized by comprising: In a destination recognition method for recognizing destination information described on a paper sheet,
A barcode detection step of detecting a barcode indicating a destination from an image on a paper sheet;
When a barcode is detected by this barcode detection process, if the barcode position is below the half position of the entire paper sheet in the longitudinal direction of the paper sheet, the barcode position is a predetermined position. A determination step of determining that the position of the barcode is valid when there is a pixel distribution that is likely to be destination information above the position of the barcode ,
If it is determined that the position of the barcode is appropriate by this determination step, a deletion step of deleting the image of the area below the barcode from the image on the paper sheet,
A recognition step of recognizing destination information from the image of the paper sheet from which the image of the lower area of the barcode has been deleted by this deletion step;
A destination recognition method comprising: In the sorting method for sorting the paper sheets based on the destination information described in the paper sheets,
A reading step of reading an image on a paper sheet conveyed one by one by a conveying means;
A barcode detection step of detecting a barcode indicating the destination from the image read by the reading step;
When a barcode is detected by this barcode detection process, if the barcode position is below the half position of the entire paper sheet in the longitudinal direction of the paper sheet, the barcode position is a predetermined position. A determination step of determining that the position of the barcode is valid when there is a pixel distribution that is likely to be destination information above the position of the barcode ,
If it is determined that the position of the barcode is appropriate by this determination step, a deletion step of deleting the image of the area below the barcode from the image on the paper sheet,
A recognition step of recognizing destination information from the image of the paper sheet from which the image of the lower area of the barcode has been deleted by this deletion step;
A sorting step of sorting paper sheets conveyed by the conveying means based on the destination information recognized by the recognition step;
A sorting method characterized by comprising:

Images