Object identification device

Abstract

This object discrimination device comprises an acquisition unit, a binarization unit, a feature extraction unit, and an area calculation unit. The acquisition unit acquires an image obtained by imaging an imaging area, in which a plurality of objects are arranged in a line, from a direction orthogonal to the arrangement direction of the objects. The binarization unit generates, from the image, a binarized image of the imaging area in which the objects are present. The feature extraction unit calculates, from the binarized image, a summed pixel value for each column of pixel values arranged in the direction orthogonal to the arrangement direction of the objects, and generates the summed pixel value columns in which the summed pixel values for the respective columns are arranged in a line in the arrangement direction of the objects. The area calculation unit calculates, on the basis of the summed pixel value columns, object boundaries in the arrangement direction of the objects in the binarized image.

Description

[Technical field]

[0001] The present invention relates to an object discrimination device, an object discrimination method, and a recording medium. [Background technology]

[0002] In a system that performs visual inspections of objects such as tablets and snacks based on images captured by a camera, a configuration is sometimes adopted in which objects to be inspected are lined up in a row and introduced one by one to an inspection device for 100% inspection (see, for example, Patent Document 1). In such an inspection system, it is necessary to calculate the object boundary in the direction of the object arrangement from an image captured by a camera of a shooting area in which multiple objects are lined up in a row and in contact with each other.

[0003] As a technique for calculating an object boundary in the direction in which a plurality of objects are arranged in a line while being in contact with one another from an image captured of a shooting area in which the objects are photographed, there is, for example, a technique described in Patent Document 2 (hereinafter, referred to as a related technique).

[0004] In the related technology, first, an X-ray is irradiated from a direction perpendicular to the alignment direction of the objects to an object to be inspected, in which a plurality of objects (such as biscuits in Patent Document 2) are vertically packed and aligned in a container, to obtain an X-ray transmission image. Next, in the related technology, a binary image of a region corresponding to the object is extracted from the acquired X-ray transmission image. Next, in the related technology, a convex region of the object in a direction perpendicular to the alignment direction of the object and perpendicular to the X-ray irradiation direction is extracted from the binary image of the extracted object region. Specifically, in the related technology, the binary image is reduced in a direction perpendicular to the alignment direction of the object, and a difference image between the reduced image and the original binary image is generated. Next, in the related technology, a convex region that is an inflection point at the top and bottom ends in a direction perpendicular to the alignment direction of the object is extracted from the difference image, and the difference image is further separated into a plurality of separation regions corresponding to the regions of each object, thereby calculating an object boundary in the alignment direction of the object. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2012-12131 A [Patent Document 2] JP 2012-242289 A Summary of the Invention [Problem to be solved by the invention]

[0006] However, the convex region extraction and separation process performed in the related technology on the binary image of the object region requires a large amount of calculation because it is performed on the image, which is two-dimensional data. Therefore, it has been difficult to calculate the object boundary in the direction of the object arrangement with a small amount of calculation from an image taken of a shooting region in which multiple objects are lined up in a row.

[0007] An object of the present invention is to provide an object discrimination device that solves the above-mentioned problems. [Means for solving the problem]

[0008] An object discrimination device according to one aspect of the present invention includes: an acquisition unit that acquires an image of a shooting area in which a plurality of objects are lined up in a row, the image being shot in a direction perpendicular to the direction in which the objects are lined up; a binarization unit that generates a binarized image of an object region, which is a region in which the object exists, from the image; a feature extraction unit that calculates a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to an arrangement direction of the objects in the binarized image, and generates a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the arrangement direction of the objects; a region calculation unit that calculates an object boundary in a direction in which the objects are aligned in the binarized image based on the sequence of total pixel values; The device is configured to include:

[0009] Further, an object discrimination method according to another aspect of the present invention includes: Acquiring an image of a photographing area in which a plurality of objects are lined up in a row, photographed from a direction perpendicular to the direction in which the objects are lined up; generating a binarized image of an object region from the image, the object being an area where the object exists; calculating a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to the alignment direction of the objects in the binarized image, and generating a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the alignment direction of the objects; calculating an object boundary in a direction in which the object is aligned in the binarized image based on the sequence of total pixel values; It is structured as follows.

[0010] In addition, a computer-readable recording medium according to another aspect of the present invention includes: On the computer, A process of acquiring an image of a photographing area in which a plurality of objects are arranged in a line, photographed from a direction perpendicular to the direction in which the objects are arranged; A process of generating a binarized image of an object region, which is a region where the object exists, from the image; a process of calculating a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to the alignment direction of the objects in the binarized image, and generating a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the alignment direction of the objects; A process of calculating an object boundary in a direction in which the object is aligned in the binarized image based on the sequence of total pixel values; The recording medium is configured to record a program for causing the recording medium to perform the above steps. Effect of the Invention

[0011] By having the above-described configuration, the present invention can calculate the object boundary in the direction in which the objects are lined up with a small amount of calculation from an image captured of a shooting area in which a plurality of objects are lined up in a row. [Brief description of the drawings]

[0012] [Figure 1] 1 is a schematic configuration diagram of an object discrimination device according to a first embodiment of the present invention. [Diagram 2] 1 is a block diagram showing an example of a processing device in an object discrimination device according to a first embodiment of the present invention. [Diagram 3]3 is a diagram showing an example of a format of image information in the object discrimination device according to the first embodiment of the present invention. FIG. [Figure 4] 3 is a diagram showing an example of a format of binarized image information in the object discrimination device according to the first embodiment of the present invention. FIG. [Diagram 5] 3 is a diagram showing an example of a format of total pixel value sequence information in the object discrimination device according to the first embodiment of the present invention. FIG. [Figure 6] 3 is a diagram showing an example of a format of object discrimination information in the object discrimination device according to the first embodiment of the present invention. FIG. [Figure 7] 3 is a diagram showing an example of a format of inspection information in the object discrimination device according to the first embodiment of the present invention. FIG. [Figure 8] 5 is a flowchart showing an example of a schematic operation of the object discrimination device according to the first embodiment of the present invention. [Figure 9] 5 is a flowchart showing an example of processing by a binarization unit in the object discrimination device according to the first embodiment of the present invention. [Figure 10] 5 is a flowchart showing an example of processing by a feature extraction unit in the object discrimination device according to the first embodiment of the present invention. [Figure 11] 2A to 2C are schematic diagrams showing an example of a binarized image, a sequence of total pixel values, a sequence of smoothed total pixel values, and a differential curve in the object discrimination device according to the first embodiment of the present invention. [Figure 12] 5 is a flowchart showing an example of processing by a region calculation unit in the object discrimination device according to the first embodiment of the present invention. [Figure 13] 1 is a schematic diagram showing an example of an object boundary (tablet boundary) in the arrangement direction of tablets and an example of an object boundary (tablet boundary) in a direction perpendicular to the arrangement direction of tablets in an object discrimination device according to a first embodiment of the present invention. FIG. [Figure 14] 5 is a flowchart showing an example of processing by an inspection unit in the object discrimination device according to the first embodiment of the present invention. [Figure 15] FIG. 11 is an explanatory diagram of an operation of a modified example of the first embodiment of the present invention. [Figure 16]FIG. 5 is a block diagram of an object discrimination device according to a second embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] [First embodiment] Next, a first embodiment of the present invention will be described in detail with reference to the drawings. In this embodiment, the object is a pharmaceutical tablet, and the present invention is applied to an object discrimination device that performs visual inspection of each tablet based on an image of a plurality of tablets captured by a camera. The type and shape of the tablet are arbitrary. The shape of the tablet may be round, capsule-shaped, disk-shaped, oval, heart-shaped, triangular rice ball-shaped, etc. Furthermore, a plurality of types of tablets with different shapes may be mixed together.

[0014] 1 is a schematic diagram of an object discrimination device 1 according to the present embodiment. Referring to FIG. 1, the object discrimination device 1 includes a camera device 10 and a processing device 20.

[0015] The camera device 10 is configured to capture a predetermined photographing area 4 set at a location where a plurality of tablets 3 flow in a line on the conveying path 2. On the upstream side of the conveying path 2, a hopper section 5 that supplies a large number of tablets and a feeder section 6 that conveys a group of tablets from the hopper section 5 to the conveying path 2 are provided. In addition, on the downstream side of the conveying path 2, a sorting section 7 that sorts the group of tablets flowing on the conveying path 2 into a group of tablets that have passed inspection and a group of tablets that have not passed inspection is provided. The conveying path 2 is configured to align the group of tablets conveyed from the feeder section 6 in a line and convey them to the sorting section 7 at a predetermined constant speed. The conveying path 2 may have a mechanism that intentionally changes the posture of the tablets 3 during conveying. The posture of the tablets 3 changes during conveying in the photographing area 4, making it possible to perform an appearance inspection of the entire circumference of the tablets 3.

[0016] The photographing direction of the camera device 10 may be set to a direction perpendicular to the arrangement direction of the tablets 3 on the conveying path 2. For example, the camera device 10 may photograph the tablets 3 flowing in a line on the conveying path 2 from directly above. The orientation of the image of the camera device 10 may be, for example, such that the direction in which the images of the tablets 3 are arranged on the image obtained by photographing coincides with the X-axis direction of the image. Hereinafter, a coordinate system is assumed in which the lower left vertex of the image of the camera device 10 is the origin O, the arrangement direction of the images of the tablets 3 in the image is the X-axis, and the direction perpendicular thereto is the Y-axis. The camera device 10 may be, for example, a high-speed color camera equipped with a CCD (Charge-Coupled Device) image sensor or a CMOS (Complementary MOS) image sensor having a pixel capacity of about several million pixels. The camera device 10 is connected to the processing device 20 by wire or wirelessly. The camera device 10 is configured to transmit the time-series images obtained by photographing to the processing device 20 together with information indicating the photographing time, etc.

[0017] The processing device 20 is an information processing device that performs image processing on the image of the photographed area 4 photographed by the camera device 10, and performs labeling processing and inspection of each tablet 3. The processing device 20 is connected to the camera device 10 and the sorting unit 7 by wire or wirelessly.

[0018] Fig. 2 is a block diagram showing an example of the processing device 20. Referring to Fig. 2, the processing device 20 includes a communication I / F unit 21, an operation input unit 22, a screen display unit 23, a storage unit 24, and a calculation processing unit 25.

[0019] The communication I / F unit 21 is composed of a data communication circuit, and is configured to perform data communication with the camera device 10, the sorting unit 7, and other external devices (not shown) by wire or wirelessly. The operation input unit 22 is composed of operation input devices such as a keyboard and a mouse, and is configured to detect an operator's operation and output it to the arithmetic processing unit 25. The screen display unit 23 is composed of a screen display device such as an LCD (Liquid Crystal Display) or a PDP (Plasma Display Panel), and is configured to display various information such as inspection results on the screen in response to instructions from the arithmetic processing unit 25.

[0020] The storage unit 24 is composed of one or more storage devices of one or more types, such as a hard disk or memory, and is configured to store processing information and a program 241 required for various processes in the arithmetic processing unit 25. The program 241 is a program that is read into the arithmetic processing unit 25 and executed to realize various processing units, and is read in advance from an external device or recording medium (not shown) via a data input / output function such as the communication I / F unit 21 and stored in the storage unit 24. Main processing information stored in the storage unit 24 includes image information 242, binarized image information 243, total pixel value string information 244, object discrimination information 245, and inspection information 246.

[0021] The image information 242 includes time-series images obtained by photographing the photographing area 4 by the camera device 10. When a plurality of tablets 3 are lined up in a line and flowing on the conveying path 2, the image information 242 shows an image of the plurality of tablets 3 lined up in a line in the X-axis direction. FIG. 3 shows an example of the format of the image information 242. The image information 242 in this example is composed of entries each consisting of a pair of a photographing time 2421 and a frame image 2422. The photographing time and the frame image are set in each of the items of the photographing time 2421 and the frame image 2422. The photographing time 2421 is set to an accuracy (for example, milliseconds) that allows the photographing time 2421 to be distinguished from other frame images.

[0022] The binary image information 243 is an image obtained by detecting and binarizing the image of the tablet 3 in the frame image 2422 of the image information 242. In this embodiment, for example, the pixel value of the image area of ​​the tablet 3 is associated with 255, and the pixel value of the other areas is associated with 0. FIG. 4 shows an example format of the binary image information 243. The binary image information 243 in this example is composed of entries each consisting of a pair of a photographing time 2431 and a binary image 2432. In each of the items of the photographing time 2431 and the binary image 2432, an image obtained by binarizing the image of the tablet 3 in the frame image 2422 and the photographing time 2421 of the image information 242 is set.

[0023] The total pixel value column information 244 is information in which the total pixel value is calculated for each column of pixel values ​​arranged in a direction (Y-axis direction) perpendicular to the arrangement direction of the tablets 3 in the binary image 2432 of the binary image information 243, and the total pixel values ​​for each column are arranged in a row in the arrangement direction of the tablets 3 (X-axis direction). FIG. 5 shows an example of the format of the total pixel value column information 244. The total pixel value column information 244 in this example is composed of entries consisting of a pair of a shooting time 2441 and a total pixel value column 2442. The shooting time 2431 of the binary image information 243 is set in the item of the shooting time 2441. The total pixel value column 2442 is set to the item of the total pixel value column 2442, which is the total pixel value 2442-1, 2442-2, ..., 2442-n for each column of pixel values ​​arranged in a direction (Y-axis direction) perpendicular to the arrangement direction of the tablets 3 in the binary image 2432 of the binary image information 243.

[0024] The object discrimination information 245 is information about tablets detected from the binary image. FIG. 6 shows an example of the format of the object discrimination information 245. The object discrimination information 245 in this example is composed of a set of an object ID 2451, a shooting time 2452, an object area 2453, and an inspection result 2454. The shooting time 2431 of the binary image information 243 is set in the shooting time 2452 item. The object area 2453 item is set with information specifying the area of ​​the image of the tablet 3 detected from the binary image 2432 at the shooting time 2431. For example, the object area 2453 is set as a rectangular area specified by the object boundary (X coordinate value) in the arrangement direction (X axis direction) of the multiple tablets in the binary image 2432 and the object boundary (Y coordinate value) in the direction perpendicular to the arrangement direction. The inspection result 2454 item is set with the result of the appearance inspection based on the image of the tablet 3 included in the object area 2453. The inspection item is arbitrary. For example, it may be at least one of a shape inspection to check for chipping of the tablet, a stain inspection to check for adhesion of foreign matter, a print inspection to check for printing errors, an engraving inspection to check for no engraving, and a color inspection to check for discoloration. When the same tablet is detected from multiple binary images 2432 with consecutive shooting times, the same object ID 2451 is assigned to the same tablet, and the sets of shooting time 2552, object area 2453, and inspection result 2454 are set as many times as the number of binary images 2432 in which the tablet is detected.

[0025] The inspection information 246 is information that indicates the total number of tablets detected from the binarized image, etc. Fig. 7 shows an example of the format of the inspection information 246. In this example, the inspection information 246 is configured to include a total number of objects 2461, a pass number 2462, and a fail number 2463. The total number of tablets detected is set in the total number of objects 2461 item. The total number of tablets that passed the inspection out of the total number of objects is set in the pass number 2462 item. The total number of tablets that did not pass the inspection out of the total number of objects is set in the fail number 2463 item.

[0026] 2 again, the arithmetic processing unit 25 has one or more microprocessors such as an MPU and its peripheral circuits, and is configured to implement various processing units by reading and executing a program 241 from the storage unit 24, thereby allowing the above hardware and the program 241 to work together. The main processing units implemented by the arithmetic processing unit 25 include an acquisition unit 251, a binarization unit 252, a feature extraction unit 253, a region calculation unit 254, and an inspection unit 255.

[0027] The acquisition unit 251 is configured to control the camera device 10 to acquire image information 242 obtained by photographing the shooting area 4 in which multiple tablets 3 flow in a line on the conveying path 2 at regular time intervals, and store the image information 242 in the memory unit 24.

[0028] The binarization unit 252 is configured to read out the image information 242 from the storage unit 24, and create a binary image 2432 of the tablet from the frame image 2422 included in the image information 242. The binarization unit 252 is also configured to create binary image information 243 including the created binary image 2432, and store it in the storage unit 24.

[0029] The feature extraction unit 253 is configured to read out the binary image information 243 from the storage unit 24, and create a total pixel value sequence 2442 from the binary image 2432 included in the binary image information 243. The feature extraction unit 253 is also configured to create total pixel value sequence information 244 including the created total pixel value sequence 2442, and store it in the storage unit 24.

[0030] The region calculation unit 254 is configured to read out the total pixel value sequence information 244 from the storage unit 24, and calculate an object region 2453 representing the region of each tablet in the binarized image 2432 based on the total pixel value sequence 2442 included in the total pixel value sequence information 244. The region calculation unit 254 is also configured to track the object region in a time-series image, and assign the same object ID to the same tablet according to the tracking result. The region calculation unit 254 is also configured to create object discrimination information 245 for each assigned object ID, and store it in the storage unit 24. The region calculation unit 254 is also configured to transmit to the inspection unit 255 the photographing time of the binarized image 2432, the object region 2453 calculated from the binarized image, and the object ID 2451 assigned to the object region 2453.

[0031] The inspection unit 255 is configured to perform an appearance inspection based on an image of a tablet to which an object ID 2451 is assigned for each object ID 2451 transmitted from the area calculation unit 254. The inspection unit 255 is also configured to record the result of the inspection in the item of the inspection result 2454 of the object discrimination information 245 corresponding to the object ID 2451. The inspection unit 255 is also configured to notify the sorting unit 7 of information (e.g., the current position of the tablet, etc.) for distinguishing and sorting tablets that have not passed the inspection from tablets that have passed the inspection. Based on the notified information, the sorting unit 7 sorts the tablet groups that have flowed through the conveying path 2 into tablet groups that have passed the inspection and tablet groups that have not passed the inspection. The inspection unit 255 is also configured to tally up the inspection results, create inspection information 246, and store it in the storage unit 24. The inspection unit 255 is also configured to display the object discrimination information 245 and the inspection information 246 on the screen display unit 23 and / or transmit them to an external device via the communication I / F unit 21.

[0032] Next, the operation of the object discrimination device 1 according to this embodiment will be described.

[0033] FIG. 8 is a flowchart showing an example of the schematic operation of the object discrimination device 1. Referring to FIG. 8, first, the acquisition unit 251 controls the camera device 10 to capture an image of the photographing area 4 in which a plurality of tablets 3 are lined up and flowing on the conveying path 2 at regular time intervals, and stores the obtained image information 242 in the storage unit 24 (step S1). Next, the binarization unit 252 reads out the image information 242 from the storage unit 24, creates a binarized image 2432 of the tablet from the frame image 2422 included in the image information 242, and stores the binarized image information 243 including the created binarized image 2432 in the storage unit 24 (step S2). Next, the feature extraction unit 253 reads out the binarized image information 243 from the storage unit 24, creates a total pixel value sequence 2442 from the binarized image 2432 included in the binarized image information 243, and stores the total pixel value sequence information 244 including the created total pixel value sequence 2442 in the storage unit 24 (step S3). Next, the region calculation unit 254 reads out the total pixel value sequence information 244 from the storage unit 24, and calculates an object region 2453 representing the region of each tablet in the binarized image 2432 based on the total pixel value sequence 2442 included in the total pixel value sequence information 244 (step S4). In addition, in step S4, the region calculation unit 254 further tracks the object region in the time-series image, assigns the same object ID to the same tablet according to the tracking result, creates object discrimination information 245 for each assigned object ID, and stores it in the storage unit 24. In addition, in step S4, the region calculation unit 254 transmits the object region 2453 calculated from the binarized image 2432 and the object ID 2451 assigned to the object region 2453 to the inspection unit 255. Next, the inspection unit 255 performs an appearance inspection based on the image of the tablet to which the object ID is assigned for each object ID 2451 transmitted from the region calculation unit 254 (step S5). In step S5, the inspection unit 255 notifies the sorting unit 7 of information for sorting the tablets. In step S5, the inspection unit 255 records the inspection result in the object discrimination information 245 in association with the object ID and the binarized image based on the inspection result. Then, the object discrimination device 1 returns to the process of step S1.Finally, the inspection unit 255 compiles the inspection results, creates inspection information 246, and stores it in the memory unit 24. It also displays the object discrimination information 245 and the inspection information 246 on the screen display unit 23 and / or transmits them to an external device via the communication I / F unit 21.

[0034] Next, the binarization unit 252 will be described in detail.

[0035] FIG. 9 is a flowchart showing an example of the detailed operation of the binarization unit 252. Referring to FIG. 9, the binarization unit 252 first converts the frame image 2422 into a grayscale image (step S11). As a result, when the frame image 2422 is an RGB image, a grayscale image in which each pixel has a value of, for example, 0 to 255 is obtained. Next, the binarization unit 252 generates a difference image between the grayscale image and a background image registered in advance (step S12). The background image is a grayscale image obtained by photographing the photographed area 4 in a state in which no tablet 3 is present. As a result, when an image of the tablet 3 is captured in the frame image 2422, a difference image in which only the image of the tablet 3 is captured is obtained. Next, the binarization unit 252 binarizes the difference image (step S13). As a result, a binarized image 2432 in which, for example, the pixel value of the image area of ​​the tablet 3 is set to 255 and the pixel value of the other areas is set to 0 is obtained.

[0036] However, the configuration of the binarization unit 252 is not limited to the above. For example, when the luminance of the area of ​​the tablet 3 is sufficiently higher than that of the background area, the binarization unit 252 may determine that an area having a luminance value higher than a preset threshold value in the grayscale image generated in step S11 is the area of ​​the tablet 3, and may binarize and extract the image of only the tablet 3.

[0037] Next, the feature extraction unit 253 will be described in detail.

[0038] Fig. 10 is a flowchart showing an example of the detailed operation of the feature extraction unit 253. Referring to Fig. 10, the feature extraction unit 253 reduces the binarized image 2432 by a predetermined magnification (step S21). This results in a binarized image having a smaller vertical and horizontal size (number of pixels) than the original binarized image 2432. The reduction magnification is determined in advance, taking into consideration the speed-up of subsequent processing and the decrease in accuracy due to reduction. However, reducing the image is not essential. The original binarized image 2432 may be treated as it is as the binarized image 2432 after reduction.

[0039] Next, the feature extraction unit 253 calculates the total pixel value for each column of pixel values ​​arranged in a direction (Y-axis direction) perpendicular to the arrangement direction of the tablets 3 in the reduced binary image 2432 (hereinafter simply referred to as the binary image 2432), and generates a total pixel value column 2442 in which the total pixel values ​​for each column are arranged in a row in the arrangement direction (X-axis direction) of the tablets 3 (step S22). Generating the total pixel value column 2442 from the total pixel values ​​for each column is also referred to as vectorization.

[0040] For example, the operation of the feature extraction unit 253 generating the total pixel value string 2442 will be described in more detail using the binarized image 2432 shown in FIG. 11 as an example. In the binarized image 2432 shown in FIG. 11, a total of four tablet images 3-1 to 3-4 are lined up in a row in the X-axis direction. Among them, the tablet image 3-1 and the tablet image 3-2 are in contact with each other. In addition, the tablet image 3-2 is in contact with another tablet 3-3 on the opposite side to the side in contact with the tablet image 3-1. In addition, there is a gap between the tablet image 3-3 and the tablet image 3-4. The feature extraction unit 253 expresses the position (x, y) of a pixel on the image using the XY coordinate system set in the binarized image 2432. Here, the origin O of the XY coordinate system is set as the lower left end point. For example, if the number of pixels in the binarized image 2432 is 700 pixels horizontally (in the X-axis direction) by 500 pixels vertically (in the Y-axis direction), the total number of columns of pixel values ​​aligned in the Y-axis direction is 700. The feature extraction unit 253 calculates the sum of the pixel values ​​of the 500 pixels that compose the column for each of these 700 columns. In the case of a column that does not cross any tablet image, all of the 500 pixels that compose the column have a pixel value of 0, so the total pixel value is 0. On the other hand, in the case of a column that crosses any tablet image, all of the pixels at the crossing points have a pixel value of 255, so the total pixel value is 255×n. Here, n is the number of pixels at the crossing points. The tablet images 3-1 to 3-4 have a rounded shape. As a result, the total number of pixels in the column that crosses the vicinity of the center of the tablet is greater than the total number of pixels in the column that crosses the vicinity of the edge of the tablet.

[0041] In Fig. 11, reference numeral 2442 indicates a graph of a total pixel value string generated by the feature extraction unit 253 from the binarized image 2432 shown in Fig. 11. The vertical axis of the total pixel value string 2442 is the total pixel value, and the horizontal axis is the X-axis (arrangement direction of the tablets 3). The total pixel value string 2442 has four peaks that correspond one-to-one to the four tablet images 3-1 to 3-4 in the binarized image 2432. In addition, valleys are formed between the peaks. The valleys are also called constrictions.

[0042] Next, the area calculation unit 254 will be described in detail.

[0043] FIG. 12 is a flowchart showing an example of the detailed operation of the region calculation unit 254. Referring to FIG. 12, the region calculation unit 254 first smoothes the total pixel value column 2442 (step S31). Smoothing the total pixel value column 2442 is an operation for reducing noise by smoothing out variations (unevenness) in the total pixel values ​​of the total pixel value column 2442. The smoothing method is not limited, and for example, moving average processing may be used. The smoothing may be repeated within a range in which information on the maximum and minimum values ​​of the total pixel value column is not lost. For example, the result of smoothing the total pixel value column 2442 may be smoothed again.

[0044] Small irregularities drawn here and there in the total pixel value sequence 2442 shown in Fig. 11 typically represent variations in the total pixel values ​​due to noise. The area calculation unit 254 reduces the effects of such noise by performing smoothing. In Fig. 11, reference numeral 2443 denotes the total pixel value sequence after the total pixel value sequence 2442 has been smoothed by the area calculation unit 254. In the total pixel value sequence 2443, the irregularities due to noise seen in the total pixel value sequence 2442 have been reduced, while the maximum and minimum values ​​of the total pixel value sequence remain.

[0045] Next, the region calculation unit 254 calculates the object boundary in the arrangement direction of the tablets in the binarized image based on the total pixel value column 2443 after smoothing. The total pixel value column 2443 shows a change in the total pixel value according to the image of the tablet present in the binarized image 2432. For example, the point where the total pixel value rises from zero in the total pixel value column 2443 almost coincides with the left end point of the tablet images 3-1 and 3-4. In addition, the point where the total pixel value falls to zero in the total pixel value column 2443 almost coincides with the right end point of the tablet images 3-3 and 3-4. In addition, the point where the total pixel value in the total pixel value column 2443 becomes the minimum almost coincides with the point where the tablet image 3-1 and the tablet image 3-2 contact each other, and the point where the tablet image 3-2 and the tablet image 3-3 contact each other. Therefore, the area calculation unit 254 calculates at least one of the points in the total pixel value column 2443 where the total pixel value rises from zero, where the total pixel value falls to zero, and where the total pixel value becomes extremely small, as the object boundary in the arrangement direction of the tablet.

[0046] In order to efficiently calculate the above-mentioned portion, in this embodiment, the region calculation unit 254 performs linear differentiation of the smoothed total pixel value column (step S32). For example, the region calculation unit 254 performs linear differentiation by taking the difference between the total pixel values ​​of two columns adjacent in the X-axis direction in the smoothed total pixel value column. Data resulting from linear differentiation of the smoothed total pixel value column is called a differential curve. The differential curve represents the magnitude of change in the total pixel value in the total pixel value column (magnitude of gradient). In FIG. 11, the reference numeral 2444 typically represents a differential curve obtained by linear differentiation of the total pixel value column 2443. The vertical axis of the differential curve represents the magnitude of gradient, and the horizontal axis represents the X-axis (arrangement direction of the tablets 3). Then, the region calculation unit 254 calculates the X-coordinate value of the rising point of the differential curve (step S33). The rising point of the differential curve is a point where the gradient changes from 0 to plus, or a point where the gradient changes from negative to 0 or positive. 11, the rising points are 2445-1, 2445-2, 2445-3, 2445-4, 2445-5, and 2445-6. These points are points in the total pixel value column 2443 where the total pixel value rises from zero, where the total pixel value falls to zero, or where the total pixel value becomes minimal.

[0047] Next, the region calculation unit 254 determines the object boundary in the X-axis direction of the tablet image on the binarized image based on the X-coordinate value of the rising point of the differential curve (step S35). This utilizes the fact that the rising point of the differential curve coincides with the end of the tablet image in the X-axis direction, as described above. For example, as shown by the dashed line in FIG. 11, the rising point 2445-1 coincides with the left end of the tablet image 3-1. Also, for example, the rising point 2445-2 coincides with the right end of the tablet image 3-1 and the left end of the tablet image 3-2. Similarly, the rising points 2445-3 to 2445-6 coincide with the ends of the tablet images 3-2 to 3-4 in the X-axis direction.

[0048] 13, object boundaries 2446-1 to 2446-6 extending in the Y-axis direction depicted on the binarized image 2432 typically represent object boundaries in the X-axis direction of the tablet image determined based on the X-coordinate values ​​of the rising points of the differential curve by the region calculation unit 254. For example, the region calculation unit 254 determines, for each rising point of the differential curve, a line that passes through the X-coordinate value of that point and is parallel to the Y-axis, as the object boundary.

[0049] Next, the region calculation unit 254 determines the object boundary in the Y-axis direction on the binarized image (step S35). Specifically, first, the region calculation unit 254 focuses on the object boundary pairs in the X-axis direction that are adjacent to each other. Next, the region calculation unit 254 determines whether or not the image of a tablet exists in the partial region on the binarized image sandwiched between the object boundary pairs under focus. If the image of a tablet does not exist, the region calculation unit 254 ends the processing for the object boundary pair under focus. If the image of a tablet exists, the region calculation unit 254 determines the pixel at the top end of the image of the tablet (the pixel of the image with the maximum Y coordinate value) as one boundary constituting the object boundary pair in the Y-axis direction. In addition, the region calculation unit 254 determines the pixel at the bottom end of the image of the tablet (the pixel of the image with the minimum Y coordinate value) as the other boundary constituting the object boundary pair in the Y-axis direction. Then, the region calculation unit 254 ends the processing for the object boundary pair under focus. When the region calculation unit 254 finishes processing the object boundary pair of interest, it determines whether or not there are any adjacent object boundary pairs in the X-axis direction that have not yet been focused on. If there are, the region calculation unit 254 shifts its attention to the remaining object boundary pairs and repeats the same processing as above. If there are no remaining, the region calculation unit 254 ends the processing of step S35.

[0050] The black circles labeled with the reference characters 2447-1 to 2447-4 in the binarized image 2432 shown in FIG.

[0051] Next, the region calculation unit 254 determines the object region (tablet region) based on the object boundary in the X-axis direction and the object boundary pair in the Y-axis direction (step S36). Specifically, first, the region calculation unit 254 focuses on the object boundary pair in the Y-axis direction. Next, the region calculation unit 254 calculates the XY coordinate values ​​of a total of four points at which a line segment passing through one boundary constituting the object boundary pair under focus and parallel to the X-axis, a line segment passing through the other boundary and parallel to the X-axis, and two object boundaries in the Y-axis direction adjacent to both sides of the object boundary pair under focus intersect. Next, the region calculation unit 254 determines a rectangle having the above four points as vertices as the object region corresponding to the object boundary pair under focus. When the region calculation unit 254 determines the object region corresponding to the object boundary pair under focus, it determines whether or not there is an unprocessed object boundary pair remaining, and if there is, it shifts its attention to one of the unprocessed object boundary pairs and repeats the same process as described above. When the area calculation section 254 has finished paying attention to all object-boundary pairs, it ends the process of step S36.

[0052] Rectangles 2448-1 to 2448-4 in the binarized image 2432 shown in FIG.

[0053] Next, the region calculation unit 254 associates the object region with the object ID (step S37). Specifically, in the case of the first acquired binary image 2432, the region calculation unit 254 assigns different object IDs to each of the object regions calculated from the binary image 2432. Then, the region calculation unit 254 creates new object discrimination information 245 having the assigned object ID 2451, a set of the photographing time 2452, the object region 2453, and the inspection result 2454 set to a NULL value, and stores it in the storage unit 24. On the other hand, in the case of the second or subsequent acquired binary image 2432, the region calculation unit 254 determines whether the object region calculated from the currently acquired binary image 2432 is due to the same tablet as the object region calculated from the binary image 2432 acquired immediately before the previous photographing time. This determination may be made, for example, by comparing the distance between, for example, the center of gravity of the object area calculated from the currently acquired binary image 2432 and the center of gravity of the object area calculated from the immediately preceding binary image 2432 with a predetermined threshold, and determining that the tablets are the same if the distance is equal to or less than the threshold. Then, the area calculation unit 254 adds a set of the photographing time 2452, the object area 2453, and the inspection result 2454 set to a NULL value of the currently acquired binary image 2432 to the object discrimination information 245 having the object ID 2451 already assigned to the same tablet, for the object area calculated from the currently acquired binary image 2432 that is the same tablet as the object area calculated from the previously acquired binary image 2432. Furthermore, the area calculation unit 254 assigns a new object ID to any object areas calculated from the currently acquired binary image 2432 that are due to a new tablet, and creates new object discrimination information 245 having the assigned object ID 2451, a set of the shooting time 2452 of the binary image 2432, the calculated object area 2453, and the inspection result 2454 set to a NULL value, and stores this in the memory unit 24.

[0054] Next, the area calculation unit 254 transmits the photographing time of the current binarized image 2432, the object area 2453 calculated therefrom, and the object ID 2451 assigned thereto to the inspection unit 255 (step S38).

[0055] Next, the inspection unit 255 will be described in detail.

[0056] FIG. 14 is a flowchart showing an example of a detailed operation of the inspection unit 255. Referring to FIG. 14, the inspection unit 255 first performs an appearance inspection (step S41). Specifically, the inspection unit 255 extracts an image portion of the object area 2453 transmitted from the area calculation unit 254 from the binarized image 2432 at the shooting time transmitted from the area calculation unit 254. Next, the inspection unit 255 analyzes the extracted image portion and performs an appearance inspection of the tablet related to the object ID transmitted from the area calculation unit 254. In this example, the inspection unit 255 performs an appearance inspection of the tablet using the binarized image, but the appearance inspection of the tablet may be performed using the frame image from which the binarized image was generated. Next, if the tablet does not pass the inspection, the inspection unit 255 notifies the sorting unit 7 of information for sorting the tablet (step S42). Next, the inspection unit 255 records the result of the inspection in the item of the inspection result 2454 of the object discrimination information 245 corresponding to the object ID (step S43). Then, the inspection unit 255 ends the process of FIG.

[0057] Furthermore, when a predetermined condition is met, such as when a request is made by an administrator, when a predetermined time arrives, or when a certain period of time has elapsed, inspection unit 255 compiles the inspection results, creates inspection information 246, and stores it in storage unit 24. Furthermore, when the above-mentioned predetermined condition is met, inspection unit 255 displays object discrimination information 245 and inspection information 246 on screen display unit 23 and / or transmits them to an external device via communication I / F unit 21.

[0058] As described above, according to this embodiment, it is possible to determine the object boundary in the arrangement direction of the tablets from an image captured of the shooting area 4 in which multiple tablets 3 are lined up in a row with a small amount of calculation. The reason for this is that the total pixel value is calculated for each column of pixel values ​​lined up in a direction perpendicular to the arrangement direction of the tablets 3 in the binarized image 2432, and the object boundary in the arrangement direction of the tablets 3 in the binarized image 2432 is determined based on the total pixel value column in which the total pixel values ​​for each column are lined up in a row in the arrangement direction of the tablets 3.

[0059] Moreover, according to this embodiment, even if a plurality of tablets 3 arranged in a row are in contact with each other in the arrangement direction, the object boundary in the arrangement direction of the tablets can be determined with high accuracy with a small amount of calculation. This is because the part where the total pixel value in the total pixel value column rises from zero, the part where the total pixel value falls to zero, and the part where the total pixel value becomes minimal are calculated as the object boundary in the arrangement direction of the tablets. Also, the rising part of the differential curve obtained by first-order differentiation of the total pixel value column is calculated as the object boundary in the arrangement direction of the tablets.

[0060] Next, a modification of this embodiment will be described. <Variation 1> The region calculation unit 254 may calculate the object boundary in the arrangement direction of the tablets in the binarized image from the result of comparing each of the total pixel values ​​in the total pixel value column with a predetermined threshold value. For example, as shown in FIG. 15, the region calculation unit 254 compares each of the total pixel values ​​in the total pixel value column 2442 shown in FIG. 11 with the threshold value TH. Then, data such as that shown by the reference numeral 300 in FIG. 15 is obtained as data representing a section in the arrangement direction of the tablets in which the total pixel value is equal to or greater than the threshold value TH. The region calculation unit 254 calculates the part in this data 300 that rises from zero and the part that falls to zero as the object boundary in the arrangement direction of the tablets. According to the first modification, the calculation accuracy of the object boundary is inferior to that of the previous embodiment, but the object boundary in the arrangement direction of the tablets can be calculated with a smaller amount of calculation. Therefore, the first modification can be sufficiently used for the purpose of counting the number of tablets.

[0061] <Variation 2> An X-ray camera such as an X-ray line sensor camera may be used instead of the camera device 10. The acquisition unit 251 may acquire a transmission X-ray image obtained by photographing an object in the photographing region 4 in which objects are lined up in a row with an X-ray camera. The binarization unit 252 may binarize pixel values, which are X-ray concentration data of the acquired transmission X-ray image. The feature extraction unit 253 may calculate a total pixel value for each row of pixel values ​​aligned in a direction perpendicular to the alignment direction of the objects in the binarized transmission X-ray image, and generate a total pixel value row in which the total pixel values ​​for each row are aligned in a row in the alignment direction of the objects. The region calculation unit 254 may calculate an object boundary in the alignment direction of the objects in the binarized transmission X-ray image based on the total pixel value row. The inspection unit 255 may perform an inspection such as counting the number of objects based on the calculated object boundary.

[0062] [Second embodiment] Next, an object discrimination device 100 according to a second embodiment of the present invention will be described with reference to Fig. 16. Fig. 16 is a block diagram of the object discrimination device 100 according to the second embodiment of the present invention.

[0063] With reference to FIG. 16, the object discrimination device 100 includes an acquisition section 110, a binarization section 120, a feature extraction section 130, and a region calculation section 140.

[0064] The acquisition unit 110 is configured to acquire an image of an imaging area in which a plurality of objects are arranged in a line, captured from a direction perpendicular to the direction in which the objects are arranged. Here, the objects may be, for example, medicines such as tablets, foods such as sweets, industrial products such as screws and bolts, etc. In addition, the three-dimensional shape of the object is arbitrary, and may be a rounded shape such as a sphere, a cylinder, an ellipse, or a disk. The acquisition unit 110 may be configured in the same manner as the acquisition unit 251 in FIG. 2, for example, but is not limited thereto.

[0065] The binarization unit 120 is configured to generate a binarized image of an object region, which is a region where an object exists, from the image acquired by the acquisition unit 110. The binarization unit 120 is configured, for example, in the same manner as the binarization unit 252 in FIG. 2, but is not limited thereto.

[0066] The feature extraction unit 130 is configured to calculate a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to the alignment direction of the objects in the binarized image generated by the binarization unit 120, and generate a total pixel value column in which the total pixel values ​​for each column are aligned in a row in the alignment direction of the objects. The feature extraction unit 130 can be configured in the same manner as the feature extraction unit 253 in FIG. 2, for example, but is not limited to this.

[0067] The region calculation unit 140 is configured to calculate an object boundary in the arrangement direction of the object in the binarized image generated by the binarization unit 120 based on the total pixel value sequence generated by the feature extraction unit 130. The region calculation unit 140 can be configured similarly to, for example, the region calculation unit 254 in FIG. 2, but is not limited thereto.

[0068] The object discrimination device 100 configured as above operates as follows. That is, the acquisition unit 110 acquires an image of a photographed area in which a plurality of objects are lined up in a row, photographed from a direction perpendicular to the direction in which the objects are lined up. Next, the binarization unit 120 generates a binarized image of an object area, which is an area in which an object exists, from the image acquired by the acquisition unit 110. Next, the feature extraction unit 130 calculates a total pixel value for each column of pixel values ​​lined up in a direction perpendicular to the direction in which the objects are lined up in the binarized image generated by the binarization unit 120, and generates a total pixel value column in which the total pixel values ​​for each column are lined up in the direction in which the objects are lined up. Next, the region calculation unit 140 calculates an object boundary in the direction in which the objects are lined up in the binarized image generated by the binarization unit 120, based on the total pixel value column generated by the feature extraction unit 130.

[0069] According to the object discrimination device 100 configured and operating as described above, it is possible to calculate the object boundary in the direction in which the objects are lined up with a small amount of calculation from an image captured in an imaging area in which a number of objects are lined up in a row. The reason for this is that the total pixel value is calculated for each column of pixel values ​​lined up in a direction perpendicular to the direction in which the objects are lined up in the binary image, a total pixel value column is generated in which the total pixel values ​​for each column are lined up in a row in the direction in which the objects are lined up, and the object boundary in the direction in which the objects are lined up in the binary image is calculated based on this total pixel value column.

[0070] Although the present invention has been described with reference to the above-mentioned embodiments, the present invention is not limited to the above-mentioned embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention. [Industrial Applicability]

[0071] The present invention can be used in fields such as counting and managing tablets and inspecting their appearance.

[0072] A part or all of the above-described embodiments can be described as, but is not limited to, the following supplementary notes. [Appendix 1] an acquisition unit that acquires an image of a shooting area in which a plurality of objects are lined up in a row, the image being shot in a direction perpendicular to the direction in which the objects are lined up; a binarization unit that generates a binarized image of an object region, which is a region in which the object exists, from the image; a feature extraction unit that calculates a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to an arrangement direction of the objects in the binarized image, and generates a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the arrangement direction of the objects; a region calculation unit that calculates an object boundary in a direction in which the objects are aligned in the binarized image based on the sequence of total pixel values; An object discrimination device comprising: [Appendix 2] the region calculation unit calculates, as the object boundary, at least one of a location where the total pixel value rises from zero, a location where the total pixel value falls to zero, and a location where the total pixel value becomes a minimum; 2. An object discrimination device according to claim 1. [Appendix 3] the region calculation unit calculates, as the object boundary, a rising point of a differential curve obtained by first-order differentiation of the sequence of total pixel values; 3. An object discrimination device according to claim 1 or 2. [Appendix 4] The feature extraction unit further performs smoothing of the sequence of total pixel values ​​before calculating the object boundary. 4. An object discrimination device according to any one of claims 1 to 3. [Appendix 5] The region calculation unit further calculates an object boundary in a direction perpendicular to a direction in which the objects are arranged in the binarized image. 5. An object discrimination device according to any one of claims 1 to 4. [Appendix 6] The object is a tablet. 6. An object discrimination device according to any one of claims 1 to 5. [Appendix 7] Acquiring an image of a photographing area in which a plurality of objects are lined up in a row, photographed from a direction perpendicular to the direction in which the objects are lined up; generating a binarized image of an object region from the image, the object being an area where the object exists; calculating a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to the alignment direction of the objects in the binarized image, and generating a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the alignment direction of the objects; calculating an object boundary in a direction in which the object is aligned in the binarized image based on the sequence of total pixel values; Object discrimination method. [Appendix 8] In the calculation of the object boundary, at least one of a point where the total pixel value rises from zero, a point where the total pixel value falls to zero, and a point where the total pixel value becomes a minimum is calculated as the object boundary. 8. The object discrimination method according to claim 7. [Appendix 9] In the calculation of the object boundary, a rising portion of a differential curve obtained by first differentiating the sequence of total pixel values ​​is calculated as the object boundary. 9. The object discrimination method according to claim 7 or 8. [Appendix 10] The calculation of the object boundary includes smoothing the sequence of total pixel values ​​before calculating the object boundary. 10. An object discrimination method according to any one of claims 7 to 9. [Appendix 11] The calculation of the object boundary further includes calculating an object boundary in a direction perpendicular to a direction in which the objects are arranged in the binarized image. 11. An object discrimination method according to any one of claims 7 to 10. [Appendix 12] The object is a tablet. 12. An object discrimination method according to any one of claims 7 to 11. [Appendix 13] On the computer, A process of acquiring an image of a photographing area in which a plurality of objects are arranged in a line, photographed from a direction perpendicular to the direction in which the objects are arranged; A process of generating a binarized image of an object region, which is a region where the object exists, from the image; a process of calculating a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to the alignment direction of the objects in the binarized image, and generating a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the alignment direction of the objects; A process of calculating an object boundary in a direction in which the object is aligned in the binarized image based on the sequence of total pixel values; A computer-readable recording medium having a program recorded thereon for causing a computer to carry out the above. [Explanation of symbols]

[0073] 1 Object discrimination device 2 Transport path 3 tablets 4. Shooting area 5 Hopper section 6 Feeder section 7 Sorting Section 10 Camera equipment 20 Processing equipment

Claims

1. an acquisition unit that acquires an image of a shooting area in which a plurality of objects are lined up in a row, the image being shot in a direction perpendicular to the direction in which the objects are lined up; a binarization unit that generates a binarized image of an object region, which is a region in which the object exists, from the image; a feature extraction unit that calculates a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to an arrangement direction of the objects in the binarized image, and generates a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the arrangement direction of the objects; a region calculation unit that calculates, based on the sequence of total pixel values, locations where the total pixel value rises from zero, locations where the total pixel value falls to zero, and locations where the total pixel value becomes minimal, as object boundaries in the arrangement direction of the objects in the binarized image, and determines, for each pair of object boundaries adjacent to each other in the arrangement direction of the objects, whether or not an image of the object is present in a partial region on the binarized image between the pair of object boundaries, and if present, determines the pair of object boundaries as the pair of object boundaries in the arrangement direction of the objects for the objects related to the image that was present; An object discrimination device comprising:

2. the region calculation unit calculates, as the object boundary, a rising point of a differential curve obtained by first-order differentiation of the sequence of total pixel values; The object discrimination device according to claim 1 .

3. The feature extraction unit further performs smoothing of the sequence of total pixel values ​​before calculating the object boundary. The object discrimination device according to claim 1 or 2.

4. The region calculation unit further calculates an object boundary in a direction perpendicular to a direction in which the objects are arranged in the binarized image.

4. An object discrimination device according to claim 1.

5. The object is a tablet.

5. An object discrimination device according to claim 1.

6. A computer comprising: Acquiring an image of a photographing area in which a plurality of objects are lined up in a row, photographed from a direction perpendicular to the direction in which the objects are lined up; generating a binarized image of an object region from the image, the object being an area where the object exists; calculating a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to the alignment direction of the objects in the binarized image, and generating a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the alignment direction of the objects; Based on the sequence of total pixel values, the locations where the total pixel value rises from zero, the locations where the total pixel value falls to zero, and the locations where the total pixel value becomes minimal are calculated as object boundaries in the arrangement direction of the objects in the binarized image, and for each pair of object boundaries adjacent to each other in the arrangement direction of the objects, it is determined whether or not an image of the object is present in a partial region on the binarized image between the pair of object boundaries, and if an image of the object is present, the pair of object boundaries is determined to be the pair of object boundaries in the arrangement direction of the objects for the objects related to the image that was present. Object discrimination method.

7. In the calculation of the object boundary, the computer calculates, as the object boundary, a rising point of a differential curve obtained by first differentiating the sequence of total pixel values. The object discrimination method according to claim 6.

8. On the computer, A process of acquiring an image of a photographing area in which a plurality of objects are arranged in a line, photographed from a direction perpendicular to the direction in which the objects are arranged; A process of generating a binarized image of an object region, which is a region where the object exists, from the image; a process of calculating a total pixel value for each column of pixel values ​​aligned in a direction perpendicular to the alignment direction of the objects in the binarized image, and generating a total pixel value column in which the total pixel values ​​for each column are aligned in a line in the alignment direction of the objects; a process of calculating, based on the sequence of total pixel values, points where the total pixel value rises from zero, points where the total pixel value falls to zero, and points where the total pixel value becomes minimal, as object boundaries in the arrangement direction of the objects in the binarized image, and for each pair of object boundaries adjacent to each other in the arrangement direction of the objects, determining whether or not an image of the object exists in a partial region on the binarized image between the pair of object boundaries, and if an image exists, determining the pair of object boundaries as the pair of object boundaries in the arrangement direction of the objects for the objects related to the image that exists; A program to carry out the above.

Images