Information processing device, information processing method, and information processing program

Abstract

An information processing device (1) comprises an abnormality inspection unit (205) and a detailed inspection unit (207). The abnormality inspection unit (205) executes a first-stage abnormality inspection for each of a plurality of inspection targets, and calculates a normality degree indicating the degree of normality for each of the plurality of inspection targets. When it is not possible to determine whether any of the plurality of inspection targets is normal or abnormal in the first-stage abnormality inspection, and all the plurality of inspection targets are designated as targets of execution of a second-stage abnormality inspection, the detailed inspection unit (207) executes the second-stage abnormality inspection preferentially for inspection targets having a relatively high degree of normality among the plurality of inspection targets on the basis of the normality degree corresponding to each of the plurality of inspection targets.

Description

Information processing device, information processing method, and information processing program

[0001] The present disclosure relates to an information processing device, an information processing method, and an information processing program.

[0002] As disclosed in Patent Document 1, in an information processing device according to the prior art that determines whether an input is normal or abnormal in two stages, when it is not possible to determine whether an input is normal or abnormal based on the determination result of a device that executes a first-stage determination process, a device that executes a second-stage determination process is used to determine whether the input is normal or abnormal again. Also, in the prior art, multiple devices that execute the first-stage determination process and multiple devices that execute the second-stage determination process are prepared depending on the situation, and the determination criteria are changed depending on the situation.

[0003] International Publication No. 2019 / 069905

[0004] In a conventional information processing device that only has one device that executes the second-stage judgment process, when multiple inputs are judged, if the judgment result of the device that executes the first-stage judgment process for a certain input cannot determine whether the input is normal or abnormal, the device that executes the second-stage judgment process unconditionally puts the processing for that input on hold if a judgment process for a previous input exists. Therefore, even if the input corresponds to an object that is relatively likely to be judged normal, it is often impossible to determine whether the input is normal or abnormal by the deadline for that input. To solve this problem, the conventional technology also has a problem that multiple devices that execute the second-stage judgment process are required. Furthermore, the conventional technology has a problem that if the judgment process result is not obtained by the deadline, subsequent processing by the information processing device becomes impossible to execute. The present disclosure aims to provide an information processing device that can relatively efficiently determine whether an input is normal or abnormal even when only one device executes the second-stage judgment process, in a technology that judges whether an input is normal or abnormal using multiple-stage judgment processes.

[0005] The information processing device according to the present disclosure comprises: an abnormality inspection unit that performs a first-stage abnormality inspection on each of a plurality of test objects and calculates a normality level indicating the degree of normality for each of the plurality of test objects; and a detailed inspection unit that, when it is not possible to determine whether any of the plurality of test objects is normal or abnormal in the first-stage abnormality inspection and all of the plurality of test objects are to be subjected to a second-stage abnormality inspection, prioritizes the second-stage abnormality inspection on test objects among the plurality of test objects that have a relatively high degree of normality, based on the normality level corresponding to each of the plurality of test objects.

[0006] According to the present disclosure, when all of the multiple inspection targets are targets for the second-stage abnormality inspection, the detailed inspection unit prioritizes the second-stage abnormality inspection for the inspection targets that have a relatively high degree of normality among the multiple inspection targets. Therefore, according to the present disclosure, in a technology for determining normality / abnormality for an input through multiple stages of determination processing, it is possible to provide an information processing device that can relatively efficiently determine whether an input is normal or abnormal even when the number of devices that executes the second-stage determination processing is only one.

[0007] 1 is a diagram showing an example of the configuration of an information processing device 1 according to embodiment 1. 2 is a diagram showing an example of the configuration of an abnormality inspection system 90 according to embodiment 1. 3 is a diagram showing an example of the configuration of a processing unit 200 according to embodiment 1. 4 is a diagram showing an example of an information list 300 according to embodiment 1. 5 is a flowchart showing the operation of the information processing device 1 according to embodiment 1. 6 is a flowchart showing the operation of the information processing device 1 according to embodiment 1. 7 is a flowchart showing the operation of a detailed inspection management unit 206 according to embodiment 1. 8 is a diagram showing an example of the hardware configuration of an information processing device 1 according to a modified example of embodiment 1. 9 is a flowchart showing the operation of an information processing device 1 according to embodiment 2. 10 is a diagram showing an example of the configuration of a processing unit 200 according to embodiment 3. 11 is a flowchart showing the operation of an information processing device 1 according to embodiment 3.

[0008] In the description of the embodiments and the drawings, the same elements and corresponding elements are given the same reference numerals. The description of elements given the same reference numerals will be omitted or simplified as appropriate. Arrows in the drawings mainly indicate the flow of data or the flow of processing. Furthermore, "unit" may be read as "circuit," "step," "procedure," "process," or "circuitry" as appropriate.

[0009] First Embodiment Hereinafter, the present embodiment will be described in detail with reference to the drawings.

[0010] *** Description of Configuration *** Fig. 1 shows an example of the configuration of an information processing device 1. The information processing device 1 has an SoC (System on Chips) 10 that executes calculations required for abnormality testing, a memory 20 that stores information for the calculations, an input interface 30, and an output interface 40. These pieces of hardware are connected as appropriate via signal lines. The information processing device 1 is, as a specific example, a general computer. The information processing device 1 may be composed of multiple computers.

[0011] The SoC 10 corresponds to a processor and has a CPU (Central Processing Unit) 11 and a GPU (Graphic Processing Unit) 12 .

[0012] The memory 20 is a combination of volatile and non-volatile storage devices. A specific example of the volatile storage device is RAM (Random Access Memory). A specific example of the non-volatile storage device is ROM (Read Only Memory), HDD (Hard Disk Drive), or flash memory.

[0013] The input interface 30 acquires data required for calculations and information for grasping timing, etc. The output interface 40 issues instructions on the timing of acquiring data required for calculations and notifies the outside of the processing results, etc. Specific examples of the input interface 30 and the output interface 40 are a USB (Universal Serial Bus) terminal, a communication chip, or a NIC (Network Interface Card).

[0014] The memory 20 stores an information processing program. The information processing program is a program that causes a computer to realize the functions of each unit of the information processing device 1. The information processing program is executed by the SoC 10. The functions of each unit of the information processing device 1 are realized by software. Data used when executing the information processing program and data obtained by executing the information processing program are stored appropriately in a storage device. Each unit of the information processing device 1 uses the storage device appropriately. As a specific example, the storage device consists of at least one of the memory 20 and a cache memory in the SoC 10. Note that the terms "data" and "information" may have the same meaning. The storage device may be independent of the computer.

[0015] The information processing program may be recorded on a computer-readable non-volatile recording medium. Specific examples of the non-volatile recording medium include an optical disk and a flash memory. The information processing program may be provided as a program product.

[0016] 2 shows an example configuration of an abnormality inspection system 90. The abnormality inspection system 90 has an information processing device 1 and a product inspection device 100. The information processing device 1 and the product inspection device 100 are connected so that they can communicate with each other. The information processing device 1 and the product inspection device 100 may be configured as an integrated device. The product inspection device 100 is a device that uses the information processing device 1 to inspect each product 101, and has a conveyor 102, a conveyor 103, a means 104, and a sensor 105.

[0017] The product 101 is an object of inspection by the product inspection device 100. A defective product 101a is a product 101 that is determined to be abnormal in an inspection by the information processing device 1. A product 101 that is not a defective product 101a is called a normal product. A normal product is a product 101 that is determined to be normal in an inspection by the information processing device 1.

[0018] The conveyor 102 is used to move the product 101 .

[0019] The conveyor 103 is used to separate the defective products 101a from the normal products.

[0020] The means 104 is a device or the like that acquires inspection data related to the product 101. A specific example of the means 104 is a camera. The inspection data is data used to inspect the product 101.

[0021] The sensor 105 is used to determine the timing for acquiring inspection data, and is specifically a laser sensor or a camera.

[0022] An arrow 106 indicates the direction in which the product 101 is moved by the conveyor 102 and the direction in which the defective product 101 a is moved by the conveyor 103 .

[0023] Distance 107 indicates the longest distance from when inspection data for a certain product 101 is acquired until it is determined whether the certain product 101 is a normal product or a defective product. A deadline time for determining whether each product 101 is a normal product or a defective product is calculated based on a value obtained by dividing the distance corresponding to distance 107 by the movement speed of each conveyor.

[0024] 3 shows an example of the configuration of a processing unit 200 that operates within the information processing device 1. The processing unit 200 has an overall control unit 201, a sensor receiving unit 202, a data receiving unit 203, an inspection control unit 204, an abnormality inspection unit 205, a detailed inspection management unit 206, a detailed inspection unit 207, and a control information transmission unit 208.

[0025] The overall control unit 201 controls the processing timing of each unit included in the processing unit 200 .

[0026] The sensor receiving unit 202 is used to determine the timing for acquiring the inspection data.

[0027] The data receiving unit 203 issues an instruction to acquire inspection data to the means 104 , and also receives the inspection data acquired by the means 104 from the product inspection device 100 .

[0028] The inspection control unit 204 controls the inspection of the product 101 in the processing unit 200 .

[0029] The anomaly inspection unit 205 performs a first-stage anomaly inspection on each of the multiple inspection objects and calculates a normality level indicating the degree of normality for each of the multiple inspection objects. The anomaly inspection is a process for determining whether or not an anomaly exists in the inspection object. As a specific example, each of the multiple inspection objects is a product inspected by the product inspection device 100. The anomaly inspection unit 205 may perform the first-stage anomaly inspection using inspection data acquired by the product inspection device 100. As a specific example, the anomaly inspection unit 205 uses the inspection data received by the data receiving unit 203 to perform an anomaly inspection on the product 101 corresponding to the inspection data.

[0030] When the first-stage abnormality inspection fails to determine whether any of the multiple inspection targets is normal or abnormal and all of the multiple inspection targets are to be subjected to the second-stage abnormality inspection, the detailed inspection management unit 206 excludes each of the multiple inspection targets for which the second-stage abnormality inspection cannot be completed by the corresponding deadline from the targets for the second-stage abnormality inspection and notifies the target that the multiple inspection targets are abnormal. Here, it is assumed that each of the multiple inspection targets has a deadline by which the corresponding abnormality inspection must be completed. The detailed inspection management unit 206 may select the inspection target for which the second-stage abnormality inspection is to be performed by referring to the information list 300. The detailed inspection management unit 206 may also use the information list 300 to identify each inspection target for which the second-stage abnormality inspection cannot be completed by the corresponding deadline. As a specific example, the detailed inspection management unit 206 manages information regarding the product 101 determined to require detailed inspection based on the inspection results by the abnormality inspection unit 205. As a specific example, when the inspection by the anomaly inspection unit 205 fails to determine whether a certain product 101 is a normal product or a defective product, it is determined that a detailed inspection is necessary for the certain product 101. The detailed inspection management unit 206 is also called a detailed anomaly inspection management unit. A detailed inspection is typically an anomaly inspection that is more detailed than the anomaly inspection by the anomaly inspection unit 205. A detailed inspection is also called a detailed anomaly inspection. A detailed inspection is a subordinate concept to an anomaly inspection.

[0031] When the detailed inspection unit 207 is unable to determine whether any of the multiple inspection objects is normal or abnormal in the first-stage abnormality inspection and all of the multiple inspection objects are to be subjected to the second-stage abnormality inspection, the detailed inspection unit 207 performs the second-stage abnormality inspection preferentially on inspection objects that have a relatively high degree of normality among the multiple inspection objects, based on the normality levels corresponding to each of the multiple inspection objects. The detailed inspection unit 207 performs the second-stage abnormality inspection on the inspection objects selected by the detailed inspection management unit 206. The detailed inspection unit 207 may perform the second-stage abnormality inspection using inspection data. As a specific example, the detailed inspection unit 207 inspects the presence or absence of abnormalities in the product 101 by performing a detailed inspection using inspection data corresponding to the product 101 determined to require detailed inspection.

[0032] The control information transmission unit 208 notifies the product inspection device 100 of the inspection results for the product 101. The inspection results for a certain product 101 may include information indicating the inspection results for the certain product 101 by the detailed inspection unit 207, in addition to information indicating the presence or absence of abnormalities in the certain product 101.

[0033] FIG. 4 shows an example of an information list 300 held in the detailed inspection management unit 206. The information list 300 includes at least an inspection target 301 and a normality level 302. Targets for the second-stage anomaly inspection are registered in the information list 300. Deadline times corresponding to each of the multiple inspection targets may also be registered in the information list 300. The inspection target 301 is information for identifying a target for which a detailed inspection is to be performed. The inspection target 301 may be information indicating the product 101, or may be information indicating inspection data corresponding to the product 101. The normality level 302 is a probability that the product is normal, and is an index calculated in the anomaly inspection by the anomaly inspection unit 205, corresponding to the probability that the corresponding product 101 is a normal product. The normality level 302 is also called an abnormality inspection result index. The normality level 302 corresponding to each of the multiple inspection targets may be registered in the information list 300.

[0034] ***Description of Operation*** The operating procedure of the information processing device 1 corresponds to an information processing method. Also, the program that realizes the operation of the information processing device 1 corresponds to an information processing program.

[0035] 5 and 6 are flowcharts showing an example of the overall operation of the information processing device 1. The overall operation of the information processing device 1 will be described with reference to FIGS.

[0036] (Step S10) When the information processing device 1 is connected to the product inspection device 100, the overall control unit 201 operating in the CPU 11 waits for the sensor receiving unit 202 to receive an output from the sensor 105 via the input interface 30.

[0037] (Step S20) When the overall control unit 201 recognizes that the sensor receiving unit 202 has received an output from the sensor 105, the process proceeds to step S30.

[0038] (Step S30) The overall control unit 201 instructs the data receiving unit 203 to acquire inspection data. The data receiving unit 203 acquires the inspection data from the means 104 via the output interface 40, and sends the acquired inspection data to the inspection control unit 204. Note that the product 101 corresponding to the inspection data acquired in this step is the target product.

[0039] (Step S40) The inspection control unit 204 receives the inspection data from the data receiving unit 203 and instructs the abnormality inspection unit 205 to perform an abnormality inspection using the received inspection data. The abnormality inspection unit 205, which operates in the CPU 11, performs an abnormality inspection using the inspection data and determines the result of the abnormality inspection. The result of the abnormality inspection is data indicating whether the target product is a normal product, requires further inspection, or is a defective product. The abnormality inspection unit 205 calculates a normality level 302 corresponding to the target product using the inspection data and determines the result of the abnormality inspection corresponding to the target product based on the calculated normality level 302.

[0040] (Step S50) If the target product is determined to be a normal product, the inspection control unit 204 proceeds to step S60. If the target product is determined to be a defective product, the inspection control unit 204 proceeds to step S70. If it is determined that a detailed inspection of the target product is necessary, the inspection control unit 204 proceeds to step S80.

[0041] (Step S60) The inspection control unit 204 passes a normal product notification indicating that the target product is a normal product to the overall control unit 201. The overall control unit 201 receives the normal product notification and instructs the control information transmission unit 208 to notify the product inspection device 100 of the received normal product notification. The control information transmission unit 208 notifies the product inspection device 100 of the normal product notification via the output interface 40.

[0042] (Step S70) The same process as in step S60 is executed for the defective product notification indicating that the target product is a defective product.

[0043] (Step S80) The same process as in step S60 is executed for the detailed inspection notification indicating that a detailed inspection of the target product is required. The process in this step is executed when it is not possible to determine whether the target product is normal or abnormal in the first-stage abnormality inspection.

[0044] (Step S90) The inspection control unit 204 prepares for a detailed inspection using the detailed inspection management unit 206. At this time, the inspection control unit 204 checks the processing status of the detailed inspection by the detailed inspection unit 207 operating in the GPU 12 using the detailed inspection management unit 206. If there is no preceding detailed inspection, the inspection control unit 204 proceeds to step S120 to immediately perform a detailed inspection. Otherwise, the inspection control unit 204 proceeds to step S100. The preceding detailed inspection is a detailed inspection that precedes the detailed inspection of the target product and is a detailed inspection that should be processed by the detailed inspection unit 207.

[0045] (Step S100) The detailed inspection management unit 206 registers information indicating the target product as the inspection target 301 in the information list 300, registers the normality level when it is determined that a detailed inspection of the target product is necessary as the normality level 302, and waits.

[0046] (Step S110) If the detailed inspection unit 207 can perform a detailed inspection of the target product without exceeding the corresponding deadline, the inspection control unit 204 proceeds to step S120. Here, a deadline is set for each detailed inspection. If it is not possible to perform a detailed inspection of the target product because, for example, the detailed inspection of the target product is expected to exceed the corresponding deadline, the detailed inspection of the target product is not performed, and the inspection control unit 204 proceeds to step S150. In this case, the detailed inspection management unit 206 excludes the target product from the targets for the second-stage abnormality inspection.

[0047] (Step S120) The detailed inspection unit 207 performs a detailed inspection of the target product using the inspection data. In this step, a second-stage abnormality inspection is performed.

[0048] (Step S130) If the detailed inspection determines that the target product is a normal product, the inspection control unit 204 proceeds to step S140. If the detailed inspection determines that the target product is a defective product, the inspection control unit 204 proceeds to step S150.

[0049] (Step S140) This step is the same as step S60. Note that the normal product notification sent in this step indicates that the results of the abnormality inspection have been updated.

[0050] (Step S150) This step is the same as step S70. Note that the defective product notification sent in this step indicates that the results of the abnormality inspection have been updated.

[0051] 7 is a flowchart showing an example of the operation of the detailed inspection management unit 206 regarding the detailed inspection. This operation will be described with reference to FIGS.

[0052] (Step S200) The detailed inspection unit 207 cannot simultaneously perform another detailed inspection while performing a prior detailed inspection. Therefore, when the detailed inspection unit 207 is performing a prior detailed inspection, the detailed inspection management unit 206 registers information for identifying the product 101 that was the subject of the detailed inspection in step S40 as the inspection target 301 in the information list 300, and registers the normality calculated in step S40 as the normality 302. In this embodiment, as a specific example, the larger the normality 302 value for a given product 101, the higher the likelihood that the given product 101 is a normal product. The information list 300 can have multiple entries. Each entry corresponds to one detailed inspection. The information list 300 stores data corresponding to each target determined to require a detailed inspection during the anomaly inspection performed in step S40 in response to inspection requests from multiple environments within the product inspection device 100. In this step, the detailed inspection management unit 206 waits for an available unit in the detailed inspection unit 207. A specific example of a state in which the detailed inspection unit 207 has free space is a state in which the detailed inspection unit 207 is not performing a detailed inspection and there is no detailed inspection in the information list 300 that the detailed inspection unit 207 should perform.

[0053] (Step S210) If the detailed inspection management unit 206 confirms that the detailed inspection unit 207 is available, the process proceeds to step S220.

[0054] (Step S220) The detailed inspection management unit 206 refers to the information list 300 and determines the detailed inspection target with a relatively high corresponding normality level 302 as the next detailed inspection target. The detailed inspection management unit 206 also simultaneously determines the deadline target. The detailed inspection management unit 206 may determine the next detailed inspection target taking into account the deadline time corresponding to each detailed inspection. Deadline targets include detailed inspections that cannot be completed before the corresponding deadline time has passed, or the products 101 corresponding to the detailed inspection. The detailed inspection unit 207 performs the detailed inspection in accordance with the determination by the detailed inspection management unit 206. In this step, when all of the multiple inspection targets are targets for performing the second-stage abnormality inspection, the inspection target is selected according to the normality level 302.

[0055] 8 is a flowchart showing an example of the operation of the detailed inspection management unit 206 regarding the deadline target determination process. This operation will be described with reference to FIG.

[0056] (Step S300) The detailed inspection management unit 206 checks each entry in the information list 300.

[0057] (Step S310) The detailed inspection management unit 206 proceeds to step S320 if there is a detailed inspection that is a deadline target in the information list 300. Various methods can be used to determine whether a detailed inspection can be completed by the deadline. One example of such a determination method is a method of determining whether a detailed inspection can be completed by the deadline. In this example, the deadline time is the time obtained by adding the available inspection time to the detection time. The detection time is the time when the sensor 105 detects the detailed inspection target. The available inspection time is a time calculated from the distance 107 and the movement speed of the product 101. The available inspection time for a given product 101 corresponds to the maximum time that can be spent inspecting the given product 101. In this example, the detailed inspection target and the detection time are separately linked. Note that the method of determining whether a detailed inspection is a deadline target is not limited to the above. When information indicating the deadline time or the like is added to the information list 300, a method based on comparing the deadline time indicated in the information list 300 with the clock of the information processing device 1 may also be used. Furthermore, the detailed inspection management unit 206 may determine whether the detailed inspection corresponding to each entry is subject to a deadline by sequentially estimating the time at which the detailed inspection corresponding to each entry in the information list 300 will be completed based on the corresponding normality level 302. At this time, the detailed inspection management unit 206 may take into consideration that each detailed inspection is subject to a deadline and may be deleted.

[0058] (Step S320) The detailed inspection management unit 206 transmits a defective product notification (update) indicating that each product 101 corresponding to the detailed inspection that is the deadline target is the deadline target.

[0059] ***Description of Effects of First Embodiment*** In this embodiment, for products 101 that cannot be determined to be either normal or defective in the first-stage abnormality inspection, detailed inspection is preferentially performed based on the normality index, which is an index indicating the likelihood of the product being normal. A detailed inspection notification is sent for the products 101 that were subject to the detailed inspection. Furthermore, a defective product notification (update) is sent for detailed inspection targets that cannot be subjected to detailed inspection due to deadline constraints. Therefore, according to this embodiment, when there are multiple detailed inspections to be performed, detailed inspections can be preferentially performed for products 101 that are relatively likely to be determined to be normal. Furthermore, for detailed inspections corresponding to deadline targets, detailed inspection notifications are sent during abnormality inspections, and defective product notifications (updates) are sent by checking using the information list 300, thereby increasing the likelihood of preventing defective products from being overlooked. Furthermore, according to this embodiment, processing can be performed relatively efficiently without preparing multiple devices for performing the second-stage judgment process.

[0060] ***Other Configurations*** <Variation 1> In the first embodiment, the larger the value of the normality 302, the more likely it is that the product is normal. However, normality 302 having different properties may be used depending on the algorithm used in the anomaly inspection unit 205. As a specific example, when the anomaly inspection unit 205 performs an anomaly inspection using an autoencoder that has learned images of normal products, the normality 302 is a value corresponding to the difference between an image of the target product and an image of the target product when it is a normal product. Therefore, in this case, the smaller the difference between the images, i.e., the smaller the value of the normality 302, the more likely it is that the corresponding product 101 is a normal product.

[0061] <Modification 2> In the first embodiment, after an abnormality inspection is performed, one of a normal product notification, a detailed inspection notification, and a defective product notification is sent. Furthermore, for the target for the detailed inspection, one of a normal product notification (updated) and a defective product notification (updated) is sent again taking into account the deadline. However, when a detailed inspection is performed, the detailed inspection notification may not be sent before the detailed inspection is performed, and one of a normal product notification and a defective product notification taking into account the deadline may be sent for the target for the detailed inspection.

[0062] <Variation 3> In step S150, in order to distinguish between a detailed inspection not being performed due to the product being subject to a deadline and a determination that the product is defective after detailed inspection, if a detailed inspection was not performed, information indicating that a detailed inspection was not performed or that the product is subject to a deadline may be transmitted.

[0063] <Modification 4> In the first embodiment, it has been described that the anomaly inspection unit 205 operates in the CPU 11, and the detail inspection unit 207 operates in the GPU 12. However, both the anomaly inspection unit 205 and the detail inspection unit 207 may operate in the CPU 11 or in the GPU 12. Furthermore, the processing device that executes the processing of the anomaly inspection unit 205 and the processing of the detail inspection unit 207 may be variable depending on the resource usage status of the CPU 11 and the GPU 12, etc.

[0064] 9 shows an example of the hardware configuration of an information processing device 1 according to this modification. The information processing device 1 includes a processing circuit 18 instead of the CPU 11, the CPU 11 and GPU 12, the CPU 11 and memory 20, the GPU 12 and memory 20, or the CPU 11, GPU 12 and memory 20. The processing circuit 18 is hardware that realizes at least a portion of the components included in the information processing device 1. The processing circuit 18 may be dedicated hardware, or may be a processor that executes a program stored in the memory 20.

[0065] When the processing circuitry 18 is dedicated hardware, the processing circuitry 18 may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. The information processing device 1 may include multiple processing circuits that replace the processing circuitry 18. The multiple processing circuits share the role of the processing circuitry 18.

[0066] In the information processing device 1, some functions may be realized by dedicated hardware, and the remaining functions may be realized by software or firmware.

[0067] The processing circuitry 18 is realized by, for example, hardware, software, firmware, or a combination of these. The CPU 11, GPU 12, memory 20, and processing circuitry 18 are collectively referred to as the "processing circuitry." In other words, the functions of the functional components of the information processing device 1 are realized by the processing circuitry. Information processing devices 1 according to other embodiments may also have a configuration similar to this modified example.

[0068] Second Embodiment The following mainly describes the differences from the above-described embodiment with reference to the drawings.

[0069] *** Description of Configuration *** The configuration of the information processing device 1 according to this embodiment is the same as the configuration of the information processing device 1 according to embodiment 1. The detailed inspection management unit 206 according to this embodiment utilizes the information list 300 when there is only one inspection target in the second-stage abnormality inspection.

[0070] In the first embodiment, when a detailed inspection becomes necessary, the detailed inspection unit 207 determines whether to immediately start the detailed inspection or to register the detailed inspection target in the information list 300, depending on whether a preceding detailed inspection is being performed. On the other hand, in the present embodiment, the detailed inspection target is always registered in the information list 300.

[0071] ***Description of Operation*** Fig. 10 is a sequence diagram showing an example of the overall operation of the information processing device 1 according to this embodiment. The overall operation of the information processing device 1 will be described with reference to Figs. 1, 2, 3, and 10. The processing from step S10 to step S80 and the processing from step S100 to step S140 or step S150 are the same as those in embodiment 1. Note that step S90 is not performed in this embodiment.

[0072] ***Explanation of Effect of Second Embodiment*** In this embodiment, after a detailed inspection notification is sent in step S80, in step S100, the detailed inspection target is always registered in the information list 300. Therefore, compared to the first embodiment, this embodiment requires overhead of registering the detailed inspection target in the information list 300 even if there is free space in the detailed inspection unit 207, but has the effect of reducing the complexity of the process when a detailed inspection is performed.

[0073] Third Embodiment The following mainly describes the differences from the above-described embodiments with reference to the drawings. In the first embodiment, the maximum number of times an abnormality test is performed is two, but the maximum number of times an abnormality test is performed may be three or more. In this embodiment, an abnormality test is performed three or more times as needed.

[0074] *** Description of Configuration *** Fig. 11 shows an example of the configuration of the processing unit 200 according to this embodiment. The detailed inspection unit 207 according to this embodiment is made up of a first detailed inspection unit 209 and a second detailed inspection unit 210. In this example, the detailed inspection unit 207 has two subordinate detailed inspection units, so the maximum number of times detailed inspection can be performed is two. However, if the detailed inspection unit 207 has an additional subordinate detailed inspection unit, the maximum number of times detailed inspection can be performed can be three or more.

[0075] The first detailed inspection unit 209 and the second detailed inspection unit 210 are similar to the detailed inspection unit 207 according to the above-described embodiment. The processing unit 200 shown in FIG. 11 can perform up to three abnormality inspections, including the abnormality inspection performed by the abnormality inspection unit 205. The first detailed inspection unit 209 performs a second-stage abnormality inspection and calculates a second normality degree indicating the degree of normality for each of the multiple inspection targets in the second-stage abnormality inspection. In a case where the first-stage abnormality inspection fails to determine whether any of the multiple inspection targets are normal or abnormal, the second-stage abnormality inspection fails to determine whether any of the multiple inspection targets are normal or abnormal, and all of the multiple inspection targets are to be subjected to the third-stage abnormality inspection, the second detailed inspection unit 210 prioritizes the third-stage abnormality inspection for inspection targets with a relatively high degree of normality based on the second normality degrees corresponding to each of the multiple inspection targets.

[0076] The detailed inspection management unit 206 according to this embodiment has an information list 300 and an information list 300a corresponding to the first detailed inspection unit 209 and the second detailed inspection unit 210, respectively, and manages each information list. The total number of information lists managed by the detailed inspection management unit 206 is equal to the total number of detailed inspection units subordinate to the detailed inspection unit 207. The information list 300a is similar to the information list 300.

[0077] ***Explanation of Operation*** Figures 12 and 13 are flowcharts showing an example of the overall operation of the information processing device 1. The overall operation of the information processing device 1 in the case where an abnormality test is performed a maximum of three times will be described with reference to Figures 1, 2, 11, 12, and 13.

[0078] (Step S200) The first detailed inspection unit 209 executes the same process as in step S120. The normality calculated in this step corresponds to the second normality.

[0079] (Step S210) This step is the same as step S50. However, if it is determined that the target product needs to be inspected again in detail, the inspection control unit 204 proceeds to step S220.

[0080] (Step S220) The inspection control unit 204 executes the same process as in step S90. However, if there is no preceding detailed inspection, the inspection control unit 204 proceeds to step S250. Otherwise, the inspection control unit 204 proceeds to step S230.

[0081] (Step S230) The detailed inspection management unit 206 executes the same process as in step S100, except that the information list 300 is replaced with an information list 300a.

[0082] (Step S240) This step is the same as step S110. However, if the detailed inspection can be performed by the second detailed inspection unit 210, the inspection control unit 204 proceeds to step S250.

[0083] (Step S250) The second detailed inspection unit 210 executes the same process as in step S120. In this step, the third stage of abnormality inspection is executed.

[0084] (Step S260) This step is the same as step S130.

[0085] ***Description of Effects of Third Embodiment*** As described above, according to this embodiment, an abnormality test can be performed three or more times as needed.

[0086] ***Other Configurations*** <Variation 6> In the third embodiment, as an example, the first detailed inspection unit 209 and the second detailed inspection unit 210 are connected in series, and the detailed inspections are performed sequentially. However, the first detailed inspection unit 209 and the second detailed inspection unit 210 may be connected in parallel, and the first detailed inspection unit 209 and the second detailed inspection unit 210 may simultaneously perform detailed inspections depending on the results of the abnormality inspection unit 205. Furthermore, either the first detailed inspection unit 209 or the second detailed inspection unit 210 may appropriately perform the detailed inspection depending on the situation.

[0087] ***Other Embodiments*** The above-described embodiments can be freely combined, or any of the components of each embodiment can be modified, or any of the components can be omitted from each embodiment. Furthermore, the embodiments are not limited to those shown in embodiments 1 to 3, and various modifications are possible as needed. The procedures described using flowcharts, etc., can be modified as appropriate.

[0088] Various aspects of the present disclosure are summarized below as appendices.

[0089] (Supplementary Note 1) An information processing device comprising: an abnormality inspection unit that performs a first-stage abnormality inspection on each of a plurality of test objects and calculates a normality level indicating the degree of normality for each of the plurality of test objects; and a detailed inspection unit that, when it is not possible to determine whether any of the plurality of test objects is normal or abnormal in the first-stage abnormality inspection and all of the plurality of test objects are to be subjected to a second-stage abnormality inspection, prioritizes the second-stage abnormality inspection on test objects among the plurality of test objects that have a relatively high degree of normality, based on the normality level corresponding to each of the plurality of test objects.

[0090] (Supplementary Note 2) A deadline is set for each of the plurality of test objects, which is the time by which the abnormality test corresponding to each of the plurality of test objects must be completed; and the information processing device according to Supplementary Note 1 further comprises: a detailed inspection management unit that, when it is not possible to determine whether any of the plurality of test objects is normal or abnormal in the first-stage abnormality test and all of the plurality of test objects are to be subjected to the second-stage abnormality test, excludes each of the plurality of test objects for which the second-stage abnormality test cannot be completed by the corresponding deadline from the subjects to be subjected to the second-stage abnormality test and notifies them that they are abnormal.

[0091] (Appendix 3) An information processing device as described in Appendix 2, wherein the normality level corresponding to each of the plurality of test objects is registered in an information list, the detailed inspection management unit refers to the information list to select the test object for which the second stage of abnormality inspection is to be performed, and the detailed inspection unit performs the second stage of abnormality inspection on the selected test object.

[0092] (Supplementary Note 4) The information processing device according to Supplementary Note 3, wherein the detailed inspection management unit utilizes the information list when there is only one inspection target in the second-stage abnormality inspection.

[0093] (Appendix 5) An information processing device according to appendix 3 or 4, wherein the information list registers a deadline time corresponding to each of the plurality of inspection objects, and the detailed inspection management unit uses the information list to identify each inspection object that cannot complete the second stage of abnormality inspection by the corresponding deadline time.

[0094] (Supplementary Note 6) The detailed inspection unit comprises a first detailed inspection unit and a second detailed inspection unit, wherein the first detailed inspection unit performs the second-stage abnormality inspection and calculates a second normality degree indicating the degree of normality for each of the plurality of inspection objects in the second-stage abnormality inspection, and when it is not possible to determine whether any of the plurality of inspection objects is normal or abnormal in the first-stage abnormality inspection, it is not possible to determine whether any of the plurality of inspection objects is normal or abnormal in the second-stage abnormality inspection, and all of the plurality of inspection objects are to be subjected to a third-stage abnormality inspection, the second detailed inspection unit performs the third-stage abnormality inspection preferentially on inspection objects among the plurality of inspection objects that have a relatively high degree of normality based on the second normality degree corresponding to each of the plurality of inspection objects. This is an information processing device described in any one of Supplementary Notes 1 to 5.

[0095] (Appendix 7) An information processing device described in any one of Appendices 1 to 6, wherein each of the plurality of inspection objects is a product to be inspected by a product inspection device, the abnormality inspection unit performs the first stage of abnormality inspection using inspection data acquired by the product inspection device, and the detailed inspection unit performs the second stage of abnormality inspection using the inspection data.

[0096] 1 Information processing device, 10 SoC, 11 CPU, 12 GPU, 20 Memory, 30 Input interface, 40 Output interface, 18 Processing circuit, 90 Abnormality inspection system, 100 Product inspection device, 101 Product, 101a Defective product, 102, 103 Conveyor, 104 Means, 105 Sensor, 106 Arrow, 107 Distance, 200 Processing unit, 201 Overall control unit, 202 Sensor receiving unit, 203 Data receiving unit, 204 Inspection control unit, 205 Abnormality inspection unit, 206 Detailed inspection management unit, 207 Detailed inspection unit, 208 Control information transmission unit, 209 First detailed inspection unit, 210 Second detailed inspection unit, 300, 300a Information list, 301 Inspection target, 302 Normality.

Claims

1. An information processing device comprising: an abnormality inspection unit that performs a first stage of abnormality testing on each of a plurality of test objects and calculates a normality level indicating the degree of normality for each of the plurality of test objects; and a detailed inspection unit that, when it is not possible to determine whether any of the plurality of test objects is normal or abnormal in the first stage of abnormality testing and all of the plurality of test objects are to be subjected to a second stage of abnormality testing, prioritizes the second stage of abnormality testing on test objects among the plurality of test objects that have a relatively high degree of normality, based on the normality level corresponding to each of the plurality of test objects.

2. A deadline is set for each of the multiple test objects, which is the time by which the abnormality test corresponding to each of the multiple test objects must be completed, and the information processing device according to claim 1 further comprises a detailed inspection management unit that, when it is not possible to determine whether any of the multiple test objects is normal or abnormal in the first-stage abnormality test and all of the multiple test objects are to be subjected to the second-stage abnormality test, excludes each of the multiple test objects for which the second-stage abnormality test cannot be completed by the corresponding deadline from the targets to be subjected to the second-stage abnormality test and notifies that it is abnormal.

3. The information processing device described in claim 2, wherein the normality level corresponding to each of the multiple inspection objects is registered in an information list, the detailed inspection management unit refers to the information list to select the inspection object for which the second stage of abnormality inspection is to be performed, and the detailed inspection unit performs the second stage of abnormality inspection on the selected inspection object.

4. The information processing device according to claim 3, wherein the detailed inspection management unit utilizes the information list when there is only one inspection target in the second stage of abnormality inspection.

5. An information processing device as described in claim 3 or 4, wherein the information list registers deadline times corresponding to each of the plurality of inspection objects, and the detailed inspection management unit uses the information list to identify each inspection object that cannot complete the second stage of abnormality inspection by the corresponding deadline time.

6. The information processing device described in any one of claims 1 to 5, wherein the detailed inspection unit comprises a first detailed inspection unit and a second detailed inspection unit, wherein the first detailed inspection unit performs the second-stage abnormality inspection and calculates a second normality degree indicating the degree of normality for each of the plurality of inspection objects in the second-stage abnormality inspection, and when it is not possible to determine whether any of the plurality of inspection objects is normal or abnormal in the first-stage abnormality inspection, it is not possible to determine whether any of the plurality of inspection objects is normal or abnormal in the second-stage abnormality inspection, and all of the plurality of inspection objects are to be subjected to a third-stage abnormality inspection, the second detailed inspection unit performs the third-stage abnormality inspection preferentially on inspection objects among the plurality of inspection objects that have a relatively high degree of normality based on the second normality degree corresponding to each of the plurality of inspection objects.

7. An information processing device according to any one of claims 1 to 6, wherein each of the plurality of inspection objects is a product to be inspected by a product inspection device, the abnormality inspection unit performs the first stage of abnormality inspection using inspection data acquired by the product inspection device, and the detailed inspection unit performs the second stage of abnormality inspection using the inspection data.

8. An information processing method in which a computer executes a first-stage abnormality test on each of a plurality of test objects, calculates a normality level indicating the degree of normality for each of the plurality of test objects, and, if the computer is unable to determine whether any of the plurality of test objects is normal or abnormal in the first-stage abnormality test and all of the plurality of test objects are to be subjected to a second-stage abnormality test, executes the second-stage abnormality test preferentially on test objects among the plurality of test objects that have a relatively high degree of normality, based on the normality level corresponding to each of the plurality of test objects.

9. An information processing program that causes an information processing device that is a computer to execute the following: an abnormality testing process that performs a first-stage abnormality test on each of a plurality of test objects and calculates a normality level indicating the degree of normality for each of the plurality of test objects; and a detailed testing process that, when it is not possible to determine whether any of the plurality of test objects is normal or abnormal in the first-stage abnormality test and all of the plurality of test objects are to be subjected to a second-stage abnormality test, prioritizes the second-stage abnormality test on test objects among the plurality of test objects that have a relatively high degree of normality based on the normality level corresponding to each of the plurality of test objects.

Images