Article inspection system, article inspection device, article inspection method, and article inspection program

Abstract

To provide an article inspection device capable of reducing erroneous determination caused by variations in imaging conditions while ensuring the accuracy of individual recognition using printing-related information.SOLUTION: An article inspection device 60 comprises: a feature point extraction unit 603 which extracts a first feature point from a first image captured by a camera 61 at the first position, and a second feature point from a second image captured by a camera 62 at the second position; a feature point comparison unit 607 which compares a third feature point extracted from a third image captured by a camera 63 with the first feature point and the second feature point; and an article inspection unit 609 which determines whether the third image is an image of the same object captured in the first and second images on the basis of the comparison results by the feature point comparison unit 607.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to an article inspection system, an article inspection device, an article inspection method, and an article inspection program, and in particular to an article inspection system, an article inspection device, an article inspection method, and an article inspection program that perform inspection using photographed images of an article. [Background technology]

[0002] As a method for individually recognizing an item, in addition to using IC tags, barcodes, and QR codes (registered trademark), a technique called artifact metrics has been proposed that uses characteristics unique to artifacts (items) to authenticate the item. Artifact metrics is disclosed, for example, in Non-Patent Document 1. The Individual Authentication System (SAMP (registered trademark)) is a technique for applying artifact metrics to printed matter printed on the packaging of items, etc. This individual authentication system uses the fact that information relating to printing and color unevenness on the surface of materials (hereinafter also referred to as "printed data") differs from product to product, to identify individual products. The individual authentication system is disclosed, for example, in Non-Patent Document 2. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Matsumoto Tsutomu, Iwashita Naoyuki, "Current Status and Issues of Information Security Technology in the Financial Sector," Bank of Japan Institute for Monetary and Economic Studies, Financial Research, April 1999, pp. 17-31. [Non-Patent Document 2] Shachihata Co., Ltd., New Business Department, SAMP Division, “Individual Authentication System SAMP”, [online], [Retrieved June 29, 2023], Internet〈https: / / www.shachihata.co.jp / products / samp / index.php〉 Summary of the Invention [Problem to be solved by the invention]

[0004] When using individual authentication to detect fraud such as resale or copying of goods, print data is generated and registered at the manufacturer. Then, print data of the goods after they have been sold is acquired and compared with the registered print data. The print data is generated by capturing an image of the printed matter using a camera. However, if the imaging conditions at the manufacturer are not strictly consistent, there is a risk of misjudgment, in which the goods do not match due to color unevenness in printing, etc., caused by the imaging conditions, even if the goods are the same.

[0005] The present invention has been made in consideration of the above points, and relates to an item inspection system, an item inspection device, an item inspection method, and an item inspection program that reduce erroneous judgments due to variations in the imaging condition while ensuring the accuracy of individual recognition using information related to printing. [Means for solving the problem]

[0006] In order to achieve the above-mentioned object, one embodiment of the present invention provides an item inspection system for inspecting an item having a printed matter on its surface, and includes a first camera that images an item placed at a first position and generates a first image, a second camera that images an item placed at a second position and generates a second image, an image acquisition unit that acquires a third image captured by the third camera, and an item inspection device that determines, based on the first image and the second image, whether the third image is an image of the same item as the above-mentioned item.

[0007] One embodiment of the present invention is an article inspection device that inspects articles having a printed matter on their surface, and includes a feature point extraction unit that extracts a first feature point from a first image captured by a first camera of an article placed at a first position, and extracts a second feature point from a second image captured by a second camera of an article placed at a second position, a feature point comparison unit that compares a third feature point extracted from a third image captured by a third camera with the first feature point and the second feature point, and an article inspection device that determines whether the third image is an image of an article identical to the above-mentioned article based on a result of the comparison by the feature point comparison unit.

[0008] One embodiment of the present invention provides an article inspection method for inspecting an article having a printed matter on its surface, and includes a first feature point extraction step of extracting a first feature point from a first image captured by a first camera of an article placed at a first position; a second feature point extraction step of extracting a second feature point from a second image captured by a second camera of an article placed at a second position; a third feature point acquisition step of acquiring a third feature point extracted from a third image captured by a third camera; a feature point comparison step of comparing the first feature point, the second feature point and the third feature point; and an article inspection step of determining whether the third image is an image of the same article as the above-mentioned article based on a result of the comparison in the feature point comparison step.

[0009] An article inspection program according to one embodiment of the present invention causes a computer to execute the above-described method. Effect of the Invention

[0010] According to the above aspects, it is possible to provide an item inspection system, an item inspection device, an item inspection method, and an item inspection program that reduce erroneous judgments due to variations in the imaging condition while ensuring the accuracy of individual recognition using information related to printing. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a diagram showing an example of inspecting items between a pharmaceutical manufacturer, a wholesaler, and a medical institution. [Diagram 2] 1 is a diagram showing a scene in which an article inspection device and an article inspection method according to an embodiment of the present invention inspect an illegally distributed article. [Diagram 3] FIG. 1 is a diagram for explaining an article inspection system according to a first embodiment. [Figure 4] 1 is a functional block diagram for explaining an article inspection device according to a first embodiment. [Diagram 5] 13A, 13B, 13C, and 13D are diagrams for explaining extraction of feature points. [Figure 6]1A and 1B are flowcharts illustrating an article inspection method according to a first embodiment. [Figure 7] 13(a), (b), and (c) are schematic diagrams for explaining the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, a first embodiment and a second embodiment of the present invention will be described with reference to the drawings. The first embodiment and the second embodiment are collectively referred to as "the present embodiment." The drawings used in the present embodiment are created for the purpose of explaining the configuration, arrangement, effects, and technical concept of the present embodiment, and do not limit the dimensions or specific shape of the invention.

[0013] [overview] Prior to a specific description of this embodiment, an overview of this embodiment will be described. Figs. 1 and 2 are diagrams for explaining situations in which this embodiment is applied. Both Figs. 1 and 2 show examples of recognizing an article using the article inspection device and article inspection method of this embodiment. However, Fig. 1 shows a situation in which the article inspection device and article inspection method of this embodiment inspect genuine articles, and Fig. 2 shows a situation in which the article inspection device and article inspection method of this embodiment inspect illegally distributed articles.

[0014] FIG. 1 shows an example of inspecting an item between a pharmaceutical manufacturer 1, a wholesaler 2, and a medical institution 3. The item shown in FIG. 1 is a pharmaceutical product manufactured by the pharmaceutical manufacturer 1 and packed in a box. The pharmaceutical manufacturer 1 captures an image of the label of the pharmaceutical product from above with a camera 11 to generate an image. The pharmaceutical manufacturer 1 extracts, for example, color unevenness (pixels with discontinuous brightness) in the generated image as a feature point. Then, at least one of the image and the feature points extracted from the image is sent to a database (DB) server 14 and stored. The pharmaceutical product is packed in a plurality of units, and the wholesaler 2 puts the product through a distribution process. At this time, a label L1 that identifies the pharmaceutical product and the pharmaceutical manufacturer may be formed on the box in which the pharmaceutical product is packed, and the image may be captured by a camera 18, and the image and the feature points extracted from the image may be stored in the DB server 14. The label L1 may be a character, symbol, or figure printed directly on the box, or a part of the character, symbol, or figure printed on a label attached to the box, or a part of the character, symbol, or figure printed on a label attached to the box.

[0015] The feature points may be stored in association with the drug's serial number, manufacturing date, pharmaceutical manufacturer, lot, etc. In this way, since the relevant information for each drug can be obtained simply by imaging the drug, the pharmaceutical manufacturer 1 does not need to attach an identification code such as a barcode to the drug.

[0016] The wholesaler 2 may generate an image by reading the label L1 of the box in which the medicine is packed with the camera 15, and may transmit the image generated by the camera 15 to the DB server 14 via the smartphone 16 or the like. The DB server 14 may receive the image from the smartphone 16, determine whether or not there are stored feature points that match the feature points of the transmitted image, and transmit data associated with the matching feature points to the smartphone 16. In this way, the wholesaler can obtain information about the pharmaceutical during the distribution process. In addition, when the characters, symbols, figures, or parts thereof printed directly on the box are used as the label L1, the pharmaceutical manufacturer 1 does not need to add a separate identifier to the box when providing the pharmaceutical information to the wholesaler 2.

[0017] The medical institution 3 purchases pharmaceuticals. Then, the camera 12 captures an image of the box of the purchased pharmaceutical and generates an image. Furthermore, the medical institution 3 may transmit the generated image to the DB server 14, for example, from the smartphone 13. The DB server 14 may extract feature points from the received image and compare them with the stored feature points. When the feature points of the image transmitted from the medical institution 3 match the stored feature points, the DB server 14 may transmit information about the pharmaceutical associated with the matching feature points to the medical institution 3. However, the system of this embodiment is not limited to this example.

[0018] For example, in this embodiment, the pharmaceutical manufacturer 1 may receive images from the wholesaler 2 and the medical institution 3 and extract feature points from the images. Then, a service may be provided in which it is determined whether the extracted feature points match feature points read from the DB server 14. In addition, in this embodiment, the wholesaler 2 and the medical institution 3 may extract feature points from images and transmit the feature points to the DB server 14. In such a case, for example, if an application for a smartphone is created to automatically extract feature points from a captured image, the wholesaler 2 and the medical institution 3 can easily extract feature points.

[0019] FIG. 2 shows a distribution process in which a formulation delivered from a pharmaceutical manufacturer 1 to a wholesaler 2 is sold at an Internet site 4 or a retailer 5. When the Internet site 4 or the retailer 5 properly buys and sells a formulation, the label of the formulation is also read by a camera 17 at the Internet site 4 or the retailer 5, and the image or feature points are transmitted to and stored in the DB server 14 together with the name of the seller. When a drug is sold at the Internet site 4 or the retailer 5, the investigation institution 6 acquires the drug by purchasing it or the like. Then, the investigation institution 6 captures an image of the label of the drug, extracts an image or feature points, and compares them with the feature points stored in the DB server 14. When the comparison result shows that the feature points of the acquired drug are not stored in the DB server 14, the investigation institution 6 may determine that the drug may not have been manufactured by the official pharmaceutical manufacturer 1. Furthermore, when the feature points of a plurality of acquired drugs match one stored feature point, the investigation institution 6 can determine that the acquired drug may be a counterfeit product. Furthermore, if the characteristic points are not recorded in the DB server during the distribution process, the investigation organization 6 may determine that there is a possibility that the medicine has been resold or handled by an illegal dealer.

[0020] [First embodiment] Next, a first embodiment will be described. In the first embodiment, an image is generated by capturing an image of a label printed on a box in which a medicine manufactured by the pharmaceutical manufacturer 1 is packaged, and feature points are extracted from the image and stored in the DB server 14. Then, an example is described in which the pharmaceutical manufacturer 1 extracts feature points from an image sent from the medical institution 3 and compares them with the feature points stored in the DB server 14.

[0021] FIG. 3 is a diagram for explaining the article inspection system of the first embodiment. An article has a printed matter on its surface. Here, the printed matter may be any material that displays characters, figures, images, etc. printed on paper or other materials, regardless of the printing technology or printing process. The printing technology may be, for example, offset printing, digital printing, flexographic printing, gravure printing, etc. The article includes not only products that are the subject of distribution, but also boxes, bags, covers, etc. in which the products are packaged. The printed matter may be on the surface of the product or the box, etc.

[0022] The article inspection system includes a camera 61, which is a first camera that captures an image of an article 30 placed at a first position P1 to generate a first image, and a second camera 62, which is a second camera that captures an image of an article 30 placed at a second position P2 to generate a second image. The conveying system also includes a conveying unit such as a belt conveyor that automatically conveys the article 30 from the first position P1 to the second position P2. In the system shown in FIG. 3, the article 30 is conveyed in the direction of the arrow in the figure, with the label L1, which is a printed matter, conveyed from the camera 61 to the camera 62. Therefore, the label L1 of the article 30 is first captured by the camera 61, and then captured by the camera 62. The article 30 in the example shown in FIG. 3 is a box in which a product is packed, and the printed matter is the label L1 attached to the box.

[0023] The item inspection system also includes an image acquisition unit that acquires a third image taken by a third camera, and an item inspection device 60 (Figure 4) that determines, based on the first image and the second image, whether the third image is an image of the same item as the item (above-mentioned item) 30 captured by cameras 61 and 62.

[0024] 4 is a functional block diagram for explaining the article inspection device 60 of the first embodiment. The article inspection device 60 includes a data transmission / reception unit 601, a feature point extraction unit 603, a feature point storage unit 605, a feature point comparison unit 607, and an article inspection unit 609.

[0025] The data transmission / reception unit 601 is used to transmit at least one of the images captured by the cameras 61 and 62 and the feature points extracted from the images to the DB server 14, and also to receive a third image captured by a third camera, the camera 63. The feature point extraction unit 603 extracts feature points from the first image, the second image, and the third image. The feature points extracted from the first image are referred to as first feature points, the feature points extracted from the second image are referred to as second feature points, and the feature points extracted from the third image are referred to as third feature points. The feature point storage unit 605 temporarily stores the first feature points and the second feature points before transmitting them to the DB server 14. In addition, when the capacity of the feature point storage unit 605 can be sufficiently secured, the first feature points and the second feature points may be stored in the feature point storage unit 605.

[0026] The feature point extraction unit 603 and the feature point comparison unit 607 compare the first and second feature points with the third feature point. The comparison is performed, for example, by comparing at least a part of the common feature points of the first and second feature points with at least a part of the third feature points (Process A). Also, for example, the comparison is performed, for example, by comparing at least a part of the first or second feature points with at least a part of the third feature points (Process B). Processes A and B will be described later.

[0027] The article inspection device 60 is realized by known computer hardware and software that causes such hardware to execute the above-mentioned functions. The hardware includes a CPU (Central Processing Unit), a memory, an interface, and the like.

[0028] In the first embodiment, the article inspection device 60 is installed in the pharmaceutical manufacturer 1. The cameras 61 and 62 are provided in the pharmaceutical manufacturer 1 together with the article inspection device 60. The camera 63 is arranged in, for example, the medical institution 3. The camera 63 may be, for example, the camera 12 described above, or a camera provided in the smartphone 13 or the like. In the first embodiment, the pharmaceutical manufacturer is provided with the article inspection device 60 that performs relatively heavy processing such as extraction and comparison of feature points, and the medical institution 3 takes an image of the label of the formulation using a smartphone or the like and transmits it to the pharmaceutical manufacturer. In this way, the burden of managing the formulation on the medical institution that purchased the formulation can be reduced. The third image may be exchanged between the pharmaceutical manufacturer 1 and the medical institution 3 via the DB server 14.

[0029] 5(a), (b), (c), and (d) are diagrams for explaining the extraction of feature points. FIG. 5(a) and FIG. 5(c) show the same label, and FIG. 5(a) is an image taken by a camera 61 (hereinafter also referred to as a "first image"), and FIG. 5(c) is an image taken by a camera 62 (hereinafter also referred to as a "second image"). FIG. 5(b) is an enlarged view of the pattern s in FIG. 5(a). FIG. 5(d) is an enlarged view of the pattern s in FIG. 5(c). Each of the black dots in FIG. 5(a) and (c) indicates a feature point. A feature point is a point where the brightness of a pixel constituting an image differs from that of the surroundings by more than a threshold value, and is generated by missing or uneven pixels. A feature point can also be generated by uneven color of the background of a pattern.

[0030] 5(a) and 5(c), although the cameras 61 and 62 capture the same label, the first image and the second image have different feature points. This occurs due to individual differences between the cameras 61 and 62, which are different cameras, and slight differences in the lighting conditions at the positions of the cameras 61 (first position) and 62 (second position), the distances between the cameras 61 and 62 and the label, and the angles between the cameras 61 and 62 and the label.

[0031] Process A The pattern s shown in FIG. 5(b) has feature points s1, s2, s3, s4, s5, s6, and s7. The pattern s shown in FIG. 5(d) has feature points s1, s3, s8, s9, and s10. Among the feature points shown in FIG. 5(d), the feature points s1 and s3 match the feature points shown in FIG. 5(b). That is, the feature points s1 and s3 are common feature points among the feature points of the first image and the feature points of the second image. In the first embodiment, the feature point comparison unit 607 extracts the third feature points from the third image and determines whether or not all or a part of the third feature points match the feature points s1 and s3. In determining whether or not the feature points match, it may be determined whether or not the two feature points to be compared are within a predetermined distance. If all or a part of the third feature points match the feature points s1 and s3, the article inspection unit 609 determines that the third image is an image obtained by capturing the same label as the first image and the second image. In determining whether the labels match, a statistical method may be used on the errors found for the positions of all the feature points being compared.

[0032] Such process A uses feature points that appear similarly even when the imaging conditions are slightly different to determine whether the labels are the same, thereby determining whether the products, such as pharmaceuticals, are the same. Therefore, process A can distinguish between labels with high accuracy by using feature points that certainly appear on the labels. However, such process A has fewer feature points that can be used for discrimination, and there is a possibility that labels may be erroneously determined to be different even when they are the same.

[0033] Process B In the first embodiment, it is determined whether or not all or a part of the third feature points match all or a part of the feature points s1 to s10. If all or a part of the third feature points match all or a part of the feature points s1 to s10, it is determined that the third image is an image captured with the same label as the first image and the second image.

[0034] This type of process B uses feature points that may or may not appear due to slight differences in imaging conditions to determine whether labels are different or not, thereby avoiding erroneous determination that identical labels are different labels due to differences in imaging conditions.

[0035] (Item Inspection Method) Figures 6(a) and (b) are flowcharts for explaining the article inspection method of the first embodiment. Figure 6(a) is a flowchart of a process for saving feature points of a first image and a second image. Figure 6(b) is a flowchart of a process for determining whether labels are the same or different, i.e., whether articles are the same or different, using the feature points saved by the process shown in Figure 6(a).

[0036] As shown in FIG. 6(a), in storing feature points, the camera 61 captures an image of a label of an article at a first position on the belt conveyor (step S801). Then, the camera 62 captures an image of the same label at a second position (step S802). The feature point extraction unit 603 extracts feature points from an image generated by capturing the image (step S803). The feature points are temporarily stored in the feature point storage unit 605, and then transmitted to the DB server 14 and stored therein (step S804). It is determined whether or not to end the inspection (step S805). If yes, the process ends (step S805: Yes), and if not, the process returns to capturing an image (step S801: No).

[0037] As shown in FIG. 6(b), in the first embodiment, a third image is acquired from the medical institution 3 (step S811). The feature point comparison unit 607 determines whether the label discrimination process is to be performed by process A or process B (step S812). When the label is to be discriminated by process A (step S812A), it is determined whether at least a part of the third feature point matches at least a part of the feature point common to the first feature point and the second feature point (step S813). If it is determined that they match (step S813: Yes), the article inspection unit 609 determines that the label imaged at the medical institution 3 matches the label imaged at the pharmaceutical manufacturer 1 (step S816). Also, in step S813, if the third feature point does not match at least a part of the feature point common to the first feature point and the second feature point (step S813: No), the article inspection unit 609 determines that the labels do not match (step S815) and ends the process.

[0038] In step S812, when the label is to be identified in process B (step S812B), it is determined whether or not at least a part of the third feature point matches at least a part of the first feature point or the second feature point (step S814). If it is determined that they match (step S814: Yes), the article inspection unit 609 determines that the label imaged at the medical institution 3 matches the label imaged at the pharmaceutical manufacturer 1 (step S816). In addition, in step S814, if the third feature point does not match at least a part of the first feature point or the second feature point (step S814: No), the article inspection unit 609 determines that the labels do not match (step S815) and ends the process.

[0039] In the first embodiment described above, the same label is imaged at different positions using different cameras, and the feature points are stored. Then, the feature points are compared with those of the label imaged in a different scene to determine whether they are the same or different. According to the first embodiment, it is possible to determine whether an article is the same or different without attaching a separate identifier to the article. In this case, the first embodiment can prevent the article from being erroneously determined to be a different article due to differences in imaging conditions, even though the article is the same.

[0040] However, the present invention is not limited to the first embodiment described above. For example, the extraction of the feature points of the third image is not limited to being executed in the article inspection device 60 of the pharmaceutical manufacturer 1, but may be performed in the medical institution 3. In such a case, as described above, an application for a smartphone may be created to automatically extract the feature points from the captured image. Then, the feature points may be transmitted from the medical institution 3 directly to the pharmaceutical manufacturer 1 or via the DB server 14. Furthermore, the first embodiment is not limited to reading out the stored feature points in the pharmaceutical manufacturer 1 and judging whether the labels are the same or different. The difference between the labels may be judged, for example, by requesting the feature points stored in the DB server 14 from the medical institution 3 and comparing them with the feature points of the third image.

[0041] In the first embodiment, the image may be stored together with the feature points in the DB server 14. Furthermore, the first embodiment is not limited to extracting pixels that are significantly different from the surrounding pixels as feature points, and it is also possible to use, for example, the distribution of brightness of the entire image as feature points. Also, for example, feature points may be determined using blurring (whiskers) that occurs when printing a pattern on paper.

[0042] Furthermore, in the first embodiment, two cameras are used, but the number of cameras is not limited to two, and more cameras may be provided to sequentially capture images of articles.

[0043] [Second embodiment] The second embodiment differs from the first embodiment in that in an object inspection system, the first position where camera 61 is placed and the second position where camera 62 is placed differ in at least one of height, angle, and lighting condition.

[0044] 7(a), (b), and (c) are schematic diagrams for explaining the second embodiment. In FIG. 7(a), a stage 70 on which cameras 61 and 62 are placed has a relatively high placement surface 701 and a relatively low placement surface 702. The camera 61 is placed on the placement surface 701, and the camera 62 is placed on the placement surface 702. In the second embodiment, the cameras 61 and 62 placed on such a stage are used to capture an image of a label attached to a box. In this case, the capturing conditions of the first image and the second image are intentionally changed, and the total number of feature points appearing in the image can be increased.

[0045] FIG. 7(b) shows an example in which the lighting conditions of the cameras 61 and 62 on the stage 71 are different. In such a case, the imaging conditions of the first image and the second image can be changed intentionally to increase the total number of feature points that appear in the images. FIG. 7(c) shows an example in which the angles of the cameras 61 and 62 on the stage 71 to the label are different. In such a case, the imaging conditions of the first image and the second image can be changed intentionally to increase the total number of feature points that appear in the images.

[0046] However, the second embodiment is not limited to changing the heights, lighting conditions, etc. of the cameras 61 and 62. In the second embodiment, it is sufficient that the heights, lighting conditions, and angles of the cameras 61 and 62 and the label are relatively different, and it is sufficient that the height, lighting conditions, and angles of the label side are different between when the camera 61 captures an image and when the camera 62 captures an image.

[0047] Next, SAMP (registered trademark) (Shachihata Authentication Management Product (SAMP (registered trademark))) applicable to the first and second embodiments described above will be described. SAMP (registered trademark) is a technology for identifying individual objects by utilizing differences in color unevenness and color loss on the surface of a print or object. Although such color unevenness is difficult to distinguish by visual inspection, it can be distinguished by using digital imaging technology. In SAMP (registered trademark), color unevenness and the like can be extracted as feature points, and the state of the feature points (distribution, scale, color) can be recorded in association with the object. The recorded data is collated with imaging data obtained by imaging the object, and it is possible to determine whether the recorded object is the same as the imaged object. With such a technology, it becomes possible to identify an object by simply photographing the object without using an IC tag or the like.

[0048] Such SAMP (registered trademark) can be applied, for example, to ensuring traceability in the manufacturing and distribution processes, preventing illegal distribution, managing medicines in hospitals and dispensing pharmacies, and guaranteeing the authenticity of documents. [Explanation of symbols]

[0049] 1. Pharmaceutical manufacturers 2. Wholesalers 3. Medical institutions 4. Internet Sites 5. Retailers 6. Research Organization 11, 12, 15, 17, 18 Camera 13, 16 Smartphone 14 Database Server 60 Article Inspection Equipment 61, 62, 63 Camera 63 Camera Stages 70 and 71 601 Data transmission and reception unit 603 Feature Point Extraction Unit 605 Feature Point Storage Unit 607 Feature point comparison section 609 Article Inspection Department 701, 702 Placement surface

Claims

1. 1. An article inspection system for inspecting an article having a printed matter on a surface thereof, comprising: a first camera configured to capture an image of the item placed at a first position to generate a first image; a second camera for capturing an image of the item placed at a second position to generate a second image; an image acquisition unit that acquires a third image captured by a third camera; and an item inspection device that determines whether the third image is an image of the same item as the item based on the first image and the second image.

2. The object inspection system of claim 1 , wherein the first camera and the second camera differ in at least one of a relative height, an angle, and a lighting condition with respect to the object.

3. The object inspection system of claim 1 , further comprising a transport section that automatically transports the object from the first location to the second location.

4. An article inspection device for inspecting an article having a printed matter on a surface thereof, comprising: a feature point extraction unit that extracts a first feature point from a first image obtained by capturing an image of the article placed at a first position with a first camera, and extracts a second feature point from a second image obtained by capturing an image of the article placed at a second position with a second camera; a feature point comparison unit that compares a third feature point extracted from a third image captured by a third camera with the first feature point and the second feature point; and an item inspection device that determines whether the third image is an image of the same item as the item based on a result of the comparison by the feature point comparison unit.

5. The item inspection device of claim 4, wherein the item inspection unit determines that the third image is an image of an item identical to the item when at least a portion of a common feature point among the first feature point and the second feature point coincides with at least a portion of the third feature point.

6. The item inspection device of claim 4, wherein the item inspection unit determines that the third image is an image of an item identical to the item when at least a portion of the first feature points or the second feature points coincide with at least a portion of the third feature points.

7. 5. The object inspection device according to claim 4, wherein the first image and the second image differ in at least one of the relative height, angle, and lighting condition between the object and the camera that captured the object.

8. 1. An article inspection method for inspecting an article having a printed matter on a surface thereof, comprising: a first feature point extraction step of extracting a first feature point from a first image captured by a first camera of the article placed at a first position; a second feature point extraction step of extracting a second feature point from a second image captured by a second camera of the article placed at a second position; a third feature point acquiring step of acquiring a third feature point extracted from a third image captured by a third camera; a feature point comparison step of comparing the first feature point, the second feature point, and the third feature point; An item inspection method comprising: an item inspection step of determining whether the third image is an image of the same item as the item based on a result of the comparison in the feature point comparison step.

9. An article inspection program that causes a computer to execute the article inspection method according to claim 8.

Images