Judgment device, judgment system, judgment method, and computer program

Abstract

To enable more accurate assessment of fruits and vegetables. [Solution] The judgment device comprises: a pass condition determination unit that determines pass conditions used for judging the produce based on related information which is information related to the number of produce items to be judged as passable; a judgment model corresponding to the pass conditions determined by the pass condition determination unit; and a judgment unit that determines whether the quality of the produce to be judged is pass or fail based on sensor information obtained for the produce to be judged.

Description

Technical Field

[0001] The present invention relates to a determination device, a determination system, a determination method, and a computer program.

Background Art

[0002] Conventionally, techniques for determining fruits and vegetables using sensors such as image sensors have been proposed. For example, Patent Document 1 discloses identifying the variety of vegetables using an image obtained by imaging the vegetables being conveyed from above.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, when making a determination regarding the quality of fruits and vegetables, if the determination is always made based on the same criteria, it may not always be the optimal determination. The present invention has been made in view of the above circumstances, and provides a technology that enables a more appropriate determination of fruits and vegetables.

Means for Solving the Problems

[0005] One aspect of the present invention includes a pass / fail condition determination unit that determines a pass / fail condition used for determining the fruits and vegetables based on related information that is information related to the number of fruits and vegetables determined to pass; a determination model according to the pass / fail condition determined by the pass / fail condition determination unit; and a determination unit that determines pass / fail regarding the quality of the fruits and vegetables to be determined based on sensor information obtained for the fruits and vegetables to be determined.

[0006] One aspect of the present invention is the above-described determination device, wherein the pass condition determination unit decides to use pass conditions that result in relatively fewer passes when the relevant information indicates that there is a high probability of more passes occurring, and decides to use pass conditions that result in relatively more passes when the relevant information indicates that there is a low probability of more passes occurring.

[0007] One aspect of the present invention is the above-described determination device, wherein the determination model is a trained model that outputs quality indicator information of the produce to be determined, and the determination unit determines whether to pass or fail based on a threshold corresponding to the pass condition and the quality indicator information.

[0008] One aspect of the present invention is the above-described determination device, wherein the determination model is a trained model that outputs pass / fail information regarding the quality of the produce to be determined, and the determination unit determines pass or fail based on the trained model according to the pass conditions.

[0009] One aspect of the present invention is a conveying system for transporting fruits and vegetables, and a determination device for determining whether the quality of the fruits and vegetables to be judged, which are transported by the conveying system, is acceptable or unacceptable. The determination device comprises a pass condition determination unit that determines pass conditions used for judging the fruits and vegetables based on related information, which is information related to the number of fruits and vegetables to be judged as acceptable, a determination model corresponding to the pass conditions determined by the pass condition determination unit, and sensor information obtained for the fruits and vegetables to be judged, and the conveying system transports fruits and vegetables that have been determined to be acceptable and fruits and vegetables that have been determined to be unacceptable to different locations.

[0010] One aspect of the present invention is a determination method comprising: a pass condition determination step in which a pass condition used for determining the produce to be judged is determined based on relevant information which is information related to the number of produce items to be judged as passable; a determination step in which a pass or fail is determined regarding the quality of the produce to be judged based on a determination model corresponding to the pass condition determined in the pass condition determination step and sensor information obtained for the produce to be judged.

[0011] One aspect of the present invention is a computer program for causing a computer to function as a judgment device, comprising: a pass condition determination unit that determines pass conditions used for judging produce based on related information which is information related to the number of produce items to be judged as passable; a judgment model corresponding to the pass conditions determined by the pass condition determination unit; and a judgment unit that determines whether the quality of the produce items to be judged is pass or fail based on sensor information obtained for the produce items to be judged. [Effects of the Invention]

[0012] This invention makes it possible to make more accurate judgments about fruits and vegetables. [Brief explanation of the drawing]

[0013] [Figure 1] This is a schematic block diagram showing the system configuration of the judgment system 100. [Figure 2] This is a diagram illustrating the schematic configuration of the transport system 80. [Figure 3] This is a diagram illustrating the schematic configuration of the transport system 80. [Figure 4] This is a diagram illustrating the schematic configuration of the transport system 80. [Figure 5] This is a schematic block diagram showing a specific example of the functional configuration of the learning device 20. [Figure 6] This flowchart shows a specific example of the processing performed by the learning device 20. [Figure 7] This is a schematic block diagram showing a specific example of the functional configuration of the determination device 30. [Figure 8] It is a flowchart showing a specific example of the processing of the determination device 30. [Figure 9] It is a diagram showing an outline of a hardware configuration example of the information processing device 90 applied to the present embodiment. [Figure 10] It is a diagram showing a modified example of the determination device 30. [Figure 11] It is a diagram for explaining the first principle for determining the second state timing information. [Figure 12] It is a diagram for explaining the second principle for determining the second state timing information. [Figure 13] It is a diagram for explaining the third principle for determining the second state timing information. [Figure 14] Regarding the third principle of the second state timing information determination, it is a diagram showing a specific example of the valve being in an open state. [Figure 15] Regarding the third principle of the second state timing information determination, it is a diagram showing a specific example of the valve being in an open state.

Mode for Carrying Out the Invention

[0014] FIG. 1 is a schematic block diagram showing the system configuration of the determination system 100. The determination system 100 includes a sensor 10, a learning device 20, a determination device 30, and a conveyance system 80. The sensor 10, the learning device 20, the determination device 30, and the conveyance system 80 are communicably connected via a network 70. The network 70 may be a network using wireless communication or a network using wired communication. The network 70 may be configured using, for example, the Internet or a local area network (LAN). The network 70 may be configured by combining a plurality of networks.

[0015] Figures 2, 3, and 4 show a schematic diagram of the configuration of the conveying system 80. Figure 2 specifically shows the configuration of the conveying system 80 when viewed from above, and Figures 3 and 4 specifically show the configuration of the conveying system 80 when viewed from the side. The conveying system 80 conveys the objects to be processed (hereinafter referred to as "objects"). Objects 60 are, for example, fruits and vegetables, and more specifically, raw materials for processed products or beverages. More specifically, object 60 are raw materials for alcoholic beverages. Specific examples of such alcoholic beverages include brewed or distilled spirits. If the raw materials are for brewed spirits, specific examples of object 60 include fruits (e.g., grapes, apples, etc.) and grains (e.g., rice, wheat, barley, etc.). If the raw materials are for distilled spirits, specific examples of object 60 include fruits (e.g., grapes, apples, etc.), grains (e.g., rice, wheat, barley, etc.) and root vegetables (potatoes, pumpkins, carrots, etc.). The object 60 described above is merely an example, and other objects 60 may be used as raw materials for brewed or distilled alcoholic beverages, for example.

[0016] The transport system 80 comprises a first transport section 81, a separation section 82, a second transport section 83, and a third transport section 84. The first transport section 81 and the second transport section 83 transport the object 60 from left to right (along the direction of arrow 801). When the object 60 reaches the right end of the first transport section 81, it moves to the separation section 82, which is provided at an angle. If it is determined that the object 60 meets a predetermined criterion, the separation section 82 is in the first state. In the first state, the separation section 82 allows the object 60 to pass over it. For example, in the first state, one or more valves constituting the separation section 82 are closed. In Figure 3, the separation section 82 is in the first state. When the separation section 82 is in the first state, the object 60 that has moved to the separation section 82 moves to the second transport section 83 via the separation section 82. The second transport section 83 and the third transport section 84 transport the object 60 to different locations.

[0017] If it is determined that the object 60 does not meet the predetermined criteria, the separation unit 82 is in the second state. In Figure 4, the separation unit 82 is in the second state. When the separation unit 82 is in the second state, the object 60 cannot pass over it. For example, in the second state, one or more valves constituting the separation unit 82 are open. When the separation unit 82 is in the second state, the object 60 that has moved to the separation unit 82 passes through the separation unit 82 and falls downward, moving to the third transport unit 84. The third transport unit 84 transports the object 60 from top to bottom in Figure 2. The object 60 that has moved to the third transport unit 84 is then subjected to necessary treatment, for example, by an operator 85. The object 60 that has been treated by the operator 85 may be returned to the first transport unit 81.

[0018] The first transport unit 81 is provided with a sensor area 11 where information about the object is acquired by the sensor 10. That is, the sensor 10 acquires information about the object located in the sensor area 11 of the first transport unit 81. The information that the sensor 10 acquires from the object may be designed to be of any type.

[0019] Sensor 10 acquires predetermined information (sensor information) about an object (e.g., fruits and vegetables). For example, sensor 10 may be configured using an imaging device. In this case, sensor 10 generates image data about the object based on electromagnetic waves of a predetermined wavelength. Sensor 10 may generate image data of the object based on visible light, infrared light, or electromagnetic waves of other wavelengths.

[0020] Sensor 10 may be a three-dimensional shape information acquisition device. In this case, sensor 10 acquires information about the shape of an object (hereinafter referred to as "shape information"). The shape information may be, for example, information about the three-dimensional shape of the object. In this case, sensor 10 may be configured using a device that measures the distance from itself to each point on the surface of the object. More specifically, sensor 10 may be configured using a measuring device that measures the distance from sensor 10 to each part of the surface of the object by irradiating the object with a laser and measuring the scattered light. A concrete example of such a sensor 10 is LIDAR (Laser Imaging Detection and Ranging). A concrete example of shape information obtained by such a sensor 10 is point cloud data.

[0021] Figure 5 is a schematic block diagram showing a specific example of the functional configuration of the learning device 20. The learning device 20 is configured using information processing equipment such as a personal computer or a server. The learning device 20 includes a communication unit 21, a storage unit 22, and a control unit 23.

[0022] The communication unit 21 is a communication device. The communication unit 21 may be configured, for example, as a network interface. The communication unit 21 communicates data with other devices via the network 70 in accordance with the control of the control unit 23. The communication unit 21 may be a wireless communication device or a wired communication device.

[0023] The storage unit 22 is configured using a storage device such as a magnetic hard disk drive or a semiconductor storage device. The storage unit 22 stores data used by the control unit 23. The storage unit 22 may also function as, for example, a training data storage unit 221 and a trained model storage unit 222.

[0024] The training data storage unit 221 stores training data used in the learning process executed in the learning device 20. The training data stored in the training data storage unit 221 includes information acquired by a sensor of the same type as the sensor 10 for the object 60 (hereinafter referred to as "sensor information") and quality index information indicating an indicator of the quality of the object 60. The quality index information may be defined by only one criterion or by multiple criteria. Specific examples of criteria for quality index information include disease, insect damage, bruising, and spoilage. When quality index information is defined by multiple criteria, quality index information obtained according to each criterion may be associated with a single piece of sensor information. For example, for a certain object, a high quality index (an indicator of good quality) may be obtained under the criterion of disease, but a low quality index (an indicator of poor quality) may be obtained under the criterion of insect damage. By obtaining quality index information independently for each criterion in this way, it becomes possible to evaluate the quality of the object 60 with greater accuracy.

[0025] The trained model storage unit 222 stores the trained model obtained by a training process using the training data stored in the training data storage unit 221.

[0026] The control unit 23 is composed of a processor such as a CPU (Central Processing Unit) and memory. The control unit 23 functions as an information control unit 231 and a learning control unit 232 when the processor executes a program. Note that all or part of the functions of the control unit 23 may be implemented using hardware such as an ASIC (Application Specific Integrated Circuit), PLD (Programmable Logic Device), or FPGA (Field Programmable Gate Array). The above program may be recorded on a computer-readable recording medium. Computer-readable recording media include, for example, portable media such as flexible disks, magneto-optical disks, ROMs, CD-ROMs, and semiconductor storage devices (e.g., SSDs: Solid State Drives), as well as storage devices such as hard disks and semiconductor storage devices built into computer systems. The above program may be transmitted via a telecommunications line.

[0027] The information control unit 231 controls the input and output of information. For example, the information control unit 231 acquires training data from other devices (information processing devices and storage media) and records it in the training data storage unit 221. For example, the information control unit 231 transmits the trained model stored in the trained model storage unit 222 to another device (for example, the judgment device 30).

[0028] The learning control unit 232 performs a learning process using the training data stored in the training data storage unit 221. Specific examples of such a learning process include, for example, supervised learning for classification such as support vector machines, random forests, or neural networks. The learning control unit 232 may, for example, generate a trained model for outputting quality index information for each criterion based on input sensor information by performing supervised learning. The learning control unit 232 generates a trained model for determining the quality of the object 60 based on input sensor information by performing supervised learning. The learning control unit 232 records the generated trained model in the trained model storage unit 222. The trained model obtained by the learning control unit 232 may be transmitted to the determination device 30 and recorded in the determination model storage unit 322 of the determination device 30. Several specific examples of trained models are described below.

[0029] (First example) The learning control unit 232 may generate a trained model for integrating each criterion to output a single quality indicator information.

[0030] (Second example) The learning control unit 232 may generate a trained model for outputting quality index information for each criterion.

[0031] (Third specific example) The learning control unit 232 determines pass / fail information for each training data based on the quality index information and pass conditions for each criterion of each training data, and assigns pass / fail information to the training data. Such training data is generated based on multiple pass conditions. For example, training data where pass / fail information is determined by pass conditions that result in more passes at a lower pass level, training data where pass / fail information is determined by pass conditions that result in fewer passes at a higher pass level, and so on. The learning control unit 232 generates a trained model for each pass condition by performing a learning process using each training data. Each trained model is a trained model for outputting pass / fail information.

[0032] (Fourth specific example) The learning control unit 232 determines the pass / fail information for each criterion in each training data based on the quality index information for each criterion of each training data and the pass conditions set for each criterion, and assigns the pass / fail information for each criterion to the training data. Such training data is generated based on multiple pass conditions set for each criterion. For example, training data in which pass / fail information is determined by pass conditions that result in more passes at a lower pass level for damage, training data in which pass / fail information is determined by pass conditions that result in fewer passes at a higher pass level for damage, and so on. The learning control unit 232 generates a trained model for each combination of criterion and pass condition by performing a learning process using each training data. Each trained model is a trained model for outputting pass / fail information.

[0033] In summary, we have described four specific examples of pre-trained models, but the configuration of a pre-trained model is not limited to any of these examples.

[0034] Figure 6 is a flowchart illustrating a specific example of the processing performed by the learning device 20. First, the information control unit 231 acquires training data (step S101). The training data may be input by a user, acquired by communication from another information device, or acquired from a recording medium connected to the learning device 20. The learning control unit 232 uses the training data to perform a learning process and records the trained model in the trained model storage unit 222 (step S102).

[0035] Figure 7 is a schematic block diagram showing a specific example of the functional configuration of the determination device 30. The determination device 30 is configured using information processing equipment such as a personal computer or a server device. The determination device 30 includes a communication unit 31, a storage unit 32, and a control unit 33.

[0036] The communication unit 31 is a communication device. The communication unit 31 may be configured, for example, as a network interface. The communication unit 31 communicates data with other devices via the network 70 in accordance with the control of the control unit 33. The communication unit 31 may be a device that performs wireless communication or a device that performs wired communication.

[0037] The storage unit 32 is configured using a storage device such as a magnetic hard disk drive or a semiconductor storage device. The storage unit 32 stores data used by the control unit 33. The storage unit 32 may also function as, for example, an associated information storage unit 321 and a decision model storage unit 322.

[0038] The related information storage unit 321 stores related information used by the pass condition determination unit 332 when performing the determination process. The related information is information related to the number of objects 60 that are determined to be passable by the judgment unit 333. The related information may also be, for example, information showing the history of judgment results by the judgment unit 333. In the history of judgment results, if there are many objects 60 that have been determined to be passable, there is a high possibility that there will be many objects 60 that will be determined to be passable in the future, and if there are few objects 60 that have been determined to be passable, there is a high possibility that there will be few objects 60 that will be determined to be passable in the future. The related information may also be, for example, information about the environment at the production site of the object. Specific examples of environmental information include information about climate (e.g., sunshine hours, rainfall, presence or absence of abnormal weather), information about soil, and geographical information.

[0039] The judgment model storage unit 322 stores the judgment model used by the judgment unit 333 when performing judgment processing. The judgment model may be configured, for example, using information from a pre-trained model generated by a learning process. Such a learning process may be performed, for example, by another device (e.g., a learning device 20) or by the device itself (judgment device 30). The judgment model does not necessarily have to be generated by a learning process. The judgment model may be configured, for example, using a lookup table that associates sensor information with quality indicator information or pass / fail information, or it may be configured in other ways.

[0040] As described above, there are four specific examples of cases where the decision model is constructed using a pre-trained model. When the decision model is constructed using a lookup table, at least two specific examples can be defined as follows.

[0041] (First example) A lookup table that integrates multiple features obtained based on sensor information to output a single pass / fail result. Such lookup tables are generated based on multiple pass / fail conditions. For example, a lookup table for determining pass / fail results based on pass / fail conditions where a lower standard results in more passes, a lookup table for determining pass / fail results based on pass / fail conditions where a higher standard results in fewer passes, and so on.

[0042] (Second example) A lookup table for outputting pass / fail information for each of the multiple features obtained based on sensor information. Such feature-specific lookup tables are generated based on multiple pass / fail conditions. For example, a lookup table for determining pass / fail information based on pass / fail conditions where more passes occur at lower standards, a lookup table for determining pass / fail information based on pass / fail conditions where fewer passes occur at higher standards, and so on.

[0043] The control unit 33 is configured using a processor such as a CPU and memory. The control unit 33 functions as an information control unit 331, a pass condition determination unit 332, and a judgment unit 333 when the processor executes a program. Note that all or part of the functions of the control unit 33 may be implemented using hardware such as an ASIC, PLD, or FPGA. The above program may be recorded on a computer-readable recording medium. Computer-readable recording media include, for example, portable media such as flexible disks, magneto-optical disks, ROMs, CD-ROMs, and semiconductor storage devices (e.g., SSDs), as well as storage devices such as hard disks and semiconductor storage devices built into computer systems. The above program may be transmitted via a telecommunications line.

[0044] The information control unit 331 acquires information from other devices such as the sensor 10 and the learning device 20. The information control unit 331 transmits information indicating the determination result obtained by the determination unit 333 to other devices such as the separation unit 82. Such information exchange between the information control unit 331 and other devices may be performed, for example, by communication via the communication unit 31.

[0045] The pass condition determination unit 332 determines the pass conditions based on the related information stored in the related information storage unit 321. The pass conditions are used in the judgment processing of the judgment unit 333. If the related information indicates that there is a relatively high probability of more passes, the pass condition determination unit 332 decides to use pass conditions that result in relatively fewer passes (pass conditions with a higher pass level). If the related information indicates that there is a relatively low probability of more passes, the pass condition determination unit 332 decides to use pass conditions that result in relatively more passes (pass conditions with a lower pass level).

[0046] If the related information indicates that a relatively large number of items have been judged as acceptable in the history of judgment results by the judgment unit 333, it is likely that a large number of items 60 will be judged as acceptable in the future. In this case, it is decided to use acceptance conditions that result in a relatively smaller number of acceptable items. If the related information indicates that a relatively smaller number of items have been judged as acceptable in the history of judgment results by the judgment unit 333, it is likely that a small number of items 60 will be judged as acceptable in the future. In this case, it is decided to use acceptance conditions that result in a relatively larger number of acceptable items.

[0047] The relevant information may include, for example, information about the environment at the production site of the object. Specific examples of environmental information include information about climate (e.g., sunshine hours, rainfall, presence or absence of abnormal weather), soil information, and geographical information. If this information indicates that higher quality objects can be obtained, it is likely that a large number of objects 60 will be judged as acceptable. If this information indicates that only lower quality objects can be obtained, it is likely that a small number of objects 60 will be judged as acceptable.

[0048] The determination unit 333 performs a determination process according to the pass conditions determined by the pass condition determination unit 332. When the determination unit 333 performs a determination process using lower pass conditions, it performs a determination process that results in more passes. When the determination unit 333 performs a determination process using higher pass conditions, it performs a determination process that results in fewer passes. The details of the processing of the determination unit 333 will be explained below for each specific example of the determination model.

[0049] (First example of a trained model) The determination unit 333 acquires one quality index information for the object 60 by using the sensor information obtained by the sensor 10 and the trained model. In this case, the acceptance condition determined by the acceptance condition determination unit 332 is an acceptance condition related to the quality index information. For example, if the quality index information is a value, the acceptance condition may be a threshold. The determination unit 333 determines whether the quality index information obtained by the trained model satisfies the acceptance condition determined by the acceptance condition determination unit 332. If the quality index information satisfies the acceptance condition, the determination unit 333 determines that the quality of the object 60 is acceptable. If the quality index information does not satisfy the acceptance condition, the determination unit 333 determines that the quality of the object 60 is unacceptable.

[0050] (Second example of a trained model) The determination unit 333 acquires quality index information for each criterion for the object 60 by using sensor information obtained by the sensor 10 and a trained model. In this case, the acceptance conditions determined by the acceptance condition determination unit 332 are the acceptance conditions for the quality index information corresponding to each criterion. The determination unit 333 determines whether the quality index information for each criterion obtained by the trained model satisfies the acceptance conditions for each criterion determined by the acceptance condition determination unit 332. The determination unit 333 determines whether the object 60 is acceptable or unacceptable based on the determination results of the quality index information for each criterion. For example, if all quality index information satisfies the acceptance conditions, the determination unit 333 may determine that the quality of the object 60 is acceptable. For example, if even one piece of quality index information does not satisfy the acceptance conditions, the determination unit 333 may determine that the quality of the object 60 is unacceptable.

[0051] (Third specific example of a trained model) When a third specific example of a trained model is used, the trained model used for the judgment process is determined according to the pass conditions determined by the pass condition determination unit 332. The judgment unit 333 obtains pass / fail information for the object 60 by using the sensor information obtained by the sensor 10 and the determined trained model. The pass / fail information indicates whether the quality of the object 60 is pass or fail.

[0052] (Fourth specific example of a trained model) When the fourth specific example of the trained model is used, the trained model used for the judgment process is determined for each criterion according to the pass conditions determined by the pass condition determination unit 332. The judgment unit 333 obtains pass / fail information for each criterion by using the sensor information obtained by the sensor 10 for the object 60 and the trained model for each determined criterion. The judgment unit 333 determines the pass / fail status of the object 60 based on the pass / fail information for each criterion. For example, if the pass / fail information for all criteria indicates a pass, the judgment unit 233 may determine that the quality of the object 60 is acceptable. For example, if even one pass / fail piece of information indicates a fail, the judgment unit 333 may determine that the quality of the object 60 is unacceptable.

[0053] (First example of a lookup table) When the first specific example of the lookup table is used, the lookup table used for the judgment process is determined according to the pass conditions determined by the pass condition determination unit 332. The judgment unit 333 obtains pass / fail information for the object 60 by using the feature quantities obtained from the sensor information of the sensor 10 and the determined lookup table. The pass / fail information indicates whether the quality of the object 60 is pass or fail.

[0054] (Second example of a lookup table) When the second specific example of the lookup table is used, the lookup table used for the judgment process is determined for each criterion according to the acceptance criteria determined by the acceptance condition determination unit 332. The judgment unit 333 obtains acceptance / rejection information for each feature quantity by using the sensor information of the sensor 10 for the object 60 and the determined lookup table for each feature quantity. The judgment unit 333 determines whether the object 60 is acceptable or unacceptable based on the acceptance / rejection information for each feature quantity. For example, if the acceptance / rejection information for all feature quantities indicates acceptance, the judgment unit 333 may determine that the quality of the object 60 is acceptable. For example, if even one acceptance / rejection piece of information indicates rejection, the judgment unit 333 may determine that the quality of the object 60 is unacceptable.

[0055] If the judgment unit 333 determines that the object is acceptable, the information control unit 331 transmits this information to the transport system 80. Upon receiving notification of acceptance, the separation unit 82 of the transport system 80 controls its state to enter the first state at the time the object 60 passes through. In this case, the object 60 reaches the second transport unit 83 and is transported by the second transport unit 83 to the acceptance item storage unit. If the judgment unit 333 determines that the object is unacceptable, the information control unit 331 transmits this information to the transport system 80. Upon receiving notification of unacceptable, the separation unit 82 of the transport system 80 controls its state to enter the second state at the time the object 60 passes through. In this case, the object 60 falls from the separation unit 82 and reaches the third transport unit 84, where it is handled by the worker 85.

[0056] Figure 8 is a flowchart illustrating a specific example of the processing performed by the determination device 30. First, the information control unit 331 acquires sensor information from the sensor 10 (step S201). The pass condition determination unit 332 determines the pass conditions based on the relevant information (step S202). The determination unit 333 uses the sensor information to perform a determination process according to the pass conditions determined by the pass condition determination unit 332 (step S203). The information control unit 331 transmits information indicating the determination result to the transport system 80 (step S204).

[0057] In the judgment system 100 configured in this way, appropriate acceptance criteria are determined based on relevant information. Then, by performing a judgment process based on the determined acceptance criteria, it becomes possible to make more appropriate judgments about the target object (e.g., fruits and vegetables). More specifically, if it is likely that a large number of acceptable products will be obtained, the acceptance criteria can be set to be stricter, making it possible to obtain a sufficient quantity of acceptable products while improving quality. On the other hand, if it is likely that the number of acceptable products will be small, the acceptance criteria can be set lower, making it possible to meet a certain standard even if the quality is not high, while preventing a shortage in the quantity of acceptable products.

[0058] Figure 9 is a schematic diagram of an example hardware configuration of an information processing device 90 applied to this embodiment. The information processing device 90 comprises a processor 91, main memory 92, communication interface 93, auxiliary storage device 94, input / output interface 95, and internal bus 96. The processor 91, main memory 92, communication interface 93, auxiliary storage device 94, and input / output interface 95 are connected to each other via the internal bus 96 so as to be able to communicate with each other. The information processing device 90 may be applied to, for example, a learning device 20 and a determination device 30. In this case, for example, the communication unit 21 and the communication unit 31 may be configured using the communication interface 93. For example, the storage unit 22 and the storage unit 32 may be configured using the auxiliary storage device 94. Furthermore, the control unit 23 and the control unit 33 may be configured using the processor 91 and the main memory 92.

[0059] (modified version) In this embodiment, the learning device 20 and the determination device 30 are configured as separate devices, but they may be configured as an integrated device. Figure 10 shows a modified example of the determination device 30 configured in this way. The storage unit 32 of the determination device 30 shown in Figure 10 also functions as a teacher data storage unit 323. The control unit 33 of the determination device 30 shown in Figure 10 also functions as a learning control unit 334. The teacher data storage unit 323 functions similarly to the teacher data storage unit 221 of the learning device 20. The learning control unit 334 functions similarly to the learning control unit 232 of the learning device 20.

[0060] The learning device 20 may be implemented using multiple information processing devices. For example, the learning device 20 may be implemented using a cloud or other device. For example, in the learning device 20, the storage unit 22 and the control unit 23 may be implemented on different information processing devices. For example, the storage unit 22 of the learning device 20 may be distributed and implemented across multiple information processing devices. The determination device 30 may be implemented using multiple information processing devices. For example, the determination device 30 may be implemented using a cloud or other device. For example, in the determination device 30, the storage unit 32 and the control unit 33 may be implemented on different information processing devices. For example, the storage unit 32 of the determination device 30 may be distributed and implemented across multiple information processing devices.

[0061] If the determination unit 333 determines that the product is unsuccessful, it may determine information regarding the timing at which the separation unit 82 enters the second state (hereinafter referred to as "second state timing information") and transmit it to the transport system 80. A specific example of the process by which the determination unit 333 determines the second state timing information will be described below.

[0062] Figure 11 is a diagram showing the configuration of the transport system 80 when viewed from the side, and is a diagram illustrating the first principle of second state timing information determination. In Figure 11, the object 60 has a part of reduced quality (hereinafter referred to as "defective part"). Hereafter, the defective part is indicated by reference numeral 61. In Figure 11, the object is being transported by the first transport unit 81 in the direction of arrow 801. Figure 11 shows the object 60a when it is located upstream of the sensor area 11, the object 60b when it is located in the center of the sensor area 11, and the object 60c when it is located upstream of the sensor area 11. Depending on the location of the defective part 61 and the state of the object 60, there are positions within the sensor area 11 where the defective part 61 is detected by the sensor 10 and positions where it is not detected.

[0063] For example, in Figure 11, the sensor 10 does not detect the defective part 61 at the location of object 60a. Similarly, the sensor 10 does not detect the defective part 61 at the location of object 60b. On the other hand, the sensor 10 may detect the defective part 61 at the location of object 60c. In the example of Figure 11, the sensor 10 detects the defective part 61 at the location of object 60c, but depending on the location and size of the defective part 61, and the shape and condition of object 60, it is also possible that the defective part 61 may be detected at the locations of object 60a or object 60b.

[0064] The determination unit 333 makes a determination based on the sensor information obtained by the sensor 10 and may determine the timing at which the separation unit 82 begins to enter the second state according to the position of the object 60 at the time it is determined to be unacceptable. Such a timing at which the object begins to enter the second state is one specific example of second state timing information. Specifically, the time it takes for the object 60 determined to be unacceptable to reach the separation unit 82 may be calculated based on the distance between the position of the object 60 and the position of the separation unit 82 at the time it is determined to be unacceptable, and the second state timing information may be determined based on that time. Basically, it is desirable that the object 60 determined to be unacceptable enters the second state at the time it reaches the separation unit 82. If this timing is too early, there is a risk that other objects (objects determined to be acceptable) that reach the separation unit 82 before the object 60 determined to be unacceptable may be unnecessarily discarded. Conversely, if this timing is too late, there is a risk that the object 60 determined to be unacceptable may not be properly discarded.

[0065] In the example in Figure 11, if object 60a is judged to be unacceptable at position Pa, the distance from position Pa to the separation unit 82 is La, so the second state timing information is determined based on the length of La and the transport speed. If object 60b is judged to be unacceptable at position Pb, the distance from position Pb to the separation unit 82 is Lb, so the second state timing information is determined based on the length of Lb and the transport speed. If object 60c is judged to be unacceptable at position Pc, the distance from position Pc to the separation unit 82 is Lc, so the second state timing information is determined based on the length of Lc and the transport speed. The further upstream the object is judged to be unacceptable, the later the second state timing information is determined, and the further downstream the object is judged to be unacceptable, the earlier the second state timing information is determined.

[0066] Figure 12 is a diagram showing the configuration of the transport system 80 when viewed from above, and illustrates the second principle of determining the second state timing information. In Figure 12, the object 60 has a defective part 61. Object 60d is transported in a long state in the direction of travel. On the other hand, object 60e is transported in a short state in the direction of travel. Therefore, the amount of time required for object 60d and object 60e from the time their downstream ends reach the separation section 82 until they fall downwards is different. The shorter the length of object 60 in the direction of travel, the shorter the time it takes to fall, and the longer the length of object 60 in the direction of travel, the longer the time required for it to fall.

[0067] Therefore, the determination unit 333 may determine the length of time for the separation unit 82 to maintain the second state based on the length of the object 60 in the direction of travel. Such a length of time is also a specific example of second state timing information. The length of the object 60 in the direction of travel may be determined, for example, based on sensor information obtained by sensor 10, or based on information obtained by other sensors. In the example in Figure 12, the length of object 60d in the direction of travel is L1, so the length of time for maintaining the second state is determined based on the length of L1. The length of object 60e in the direction of travel is L2, so the length of time for maintaining the second state is determined based on the length of L2. The second state timing information is determined such that longer objects 60 maintain the second state for a longer period of time, and shorter objects 60 maintain the second state for a shorter period of time.

[0068] In Figure 11, one sensor is shown as an example of sensor 10, but multiple sensors 10 may be provided with different installation positions and angles. By providing sensors 10 in this way, it becomes possible to detect the defective part 61 with higher accuracy when the defective part 61 can only be detected from a specific angle as described above. In this case, the determination unit 333 may determine the position of the object 60 based, for example, on the position of the sensor 10 used in the determination process.

[0069] Figure 13 is a diagram showing the configuration of the transport system 80 when viewed from above, and illustrates the third principle of second state timing information determination. In Figure 13, the object 60 has a defective part 61. Object 60d is transported in a state that is short in the width direction (the direction perpendicular to the direction of travel in the horizontal plane). The width is W1. On the other hand, object 60e is transported in a state that is long in the width direction. The width is W2. Therefore, the number of valves that object 60d and object 60e pass through when falling in the separation section 82 is different. In the separation section 82, multiple valves are arranged in the width direction, and objects that are long in the width direction pass through more valves, while objects that are short in the width direction pass through fewer valves.

[0070] In the second state, all valves may always be opened, but in that case, if objects 60 judged to be unacceptable and objects 60 judged to be acceptable are transported side by side in the width direction, there is a risk that objects 60 judged to be acceptable may also be unnecessarily discarded. Therefore, the second state timing information relating to the third principle includes information on the number and position of valves that open when the second state is reached. The determination unit 333 determines information on the number and position of valves that open when the second state is reached according to the width direction length of the object 60 judged to be unacceptable, and includes it in the second state timing information. The determination unit 333 determines the position of the object judged to be unacceptable in the width direction based on sensor information such as image data. For example, when a coordinate system with axes in the width direction is provided, it is determined that there is an object 60d that is unacceptable between coordinates P2 and P3, and there is an object 60e that is unacceptable between coordinates P1 and P4. These coordinates may be included in the second state timing information as information on the position of the valves that open. A width direction coordinate may be associated with each valve. The determination unit 333 determines the valves associated with each of the two coordinates in the width direction of the object 60 that fails the inspection, and may determine to open those two valves and all the valves between them, and include this information in the second state timing information.

[0071] Figure 14 shows a specific example of the valve open state in relation to the third principle of second state timing information determination. In Figure 14, the valves corresponding to the position of the object 60e in the width direction are open. Specifically, of the valves arranged in the width direction, all valves from the top to the sixth valve in Figure 14 are open. By opening in this manner, it becomes possible to properly dispose of the object 60e.

[0072] Figure 15 shows a specific example of a valve in an open state, relating to the third principle of determining the second state timing information. In Figure 15, valves corresponding to the position of object 60d in the width direction are open. Specifically, of the valves lined up in the width direction, the third and fourth valves from the top in Figure 15 are open. By opening in this way, it is possible to appropriately dispose of object 60d, while controlling the valves so that other objects 60 that are close to object 60d (lined up in the width direction) when viewed along the axis of travel is prevented from being discarded.

[0073] The three principles of second-state timing information have been explained above, but it is also possible for multiple principles to be combined to define second-state timing information.

[0074] Although embodiments of this invention have been described in detail above with reference to the drawings, the specific configuration is not limited to these embodiments and includes designs and the like that do not depart from the spirit of this invention. [Explanation of Symbols]

[0075] 100…Judgment system, 20…Learning device, 21…Communication unit, 22…Storage unit, 221…Teacher data storage unit, 222…Trained model storage unit, 23…Control unit, 231…Information control unit, 232…Learning control unit, 30…Judgment device, 31…Communication unit, 32…Storage unit, 321…Related information storage unit, 322…Judgment model storage unit, 33…Control unit, 331…Information control unit, 332…Pass condition determination unit, 333…Judgment unit, 60…Object, 80…Transportation system, 81…First transport unit, 82…Separation unit, 83…Second transport unit, 84…Third transport unit

Claims

1. A pass condition determination unit determines the pass conditions used for determining the fruits and vegetables to be judged, based on related information which is information related to the number of fruits and vegetables that are judged to be passable. A determination unit determines whether the quality of the produce subject to determination is pass or fail based on a determination model corresponding to the pass conditions determined by the pass condition determination unit and sensor information obtained for the produce subject to determination, A determination device equipped with the following features.

2. The determination device according to claim 1, wherein the pass condition determination unit decides to use pass conditions that result in relatively fewer passes when the relevant information indicates that there is a high probability of more passes occurring, and decides to use pass conditions that result in relatively more passes when the relevant information indicates that there is a low probability of more passes occurring.

3. The determination device according to claim 1 or 2, wherein the determination model is a trained model that outputs quality indicator information of the produce to be determined, and the determination unit determines whether to pass or fail based on a threshold corresponding to the pass condition and the quality indicator information.

4. The determination device according to claim 1 or 2, wherein the determination model is a trained model that outputs pass / fail information regarding the quality of the produce to be determined, and the determination unit determines pass or fail based on the trained model corresponding to the pass conditions.

5. A transport system for transporting fruits and vegetables, The system includes a determination device that determines whether the quality of the produce to be judged, which is transported by the aforementioned transport system, is acceptable or unacceptable. The determination device includes a pass condition determination unit that determines the pass conditions used for determining the fruits and vegetables, based on related information which is information related to the number of fruits and vegetables that are determined to pass the determination, The system comprises a determination unit that determines whether the quality of the produce subject to evaluation is pass or fail based on a determination model corresponding to the pass conditions determined by the pass condition determination unit and sensor information obtained for the produce subject to evaluation, The aforementioned transport system is a judgment system that transports fruits and vegetables that have been judged to be acceptable to different locations from those that have been judged to be unacceptable.

6. A pass condition determination step in which the pass conditions used for determining the pass conditions for the said produce are determined based on related information, which is information related to the number of fruits and vegetables that are judged to be passable for the produce subject to evaluation, A determination step in which the quality of the produce to be judged is determined to pass or fail based on a determination model corresponding to the pass conditions determined in the pass condition determination step and sensor information obtained for the produce to be judged, A method for determining the thumbnail.

7. A pass condition determination unit determines the pass conditions used for determining the fruits and vegetables to be judged, based on related information which is information related to the number of fruits and vegetables that are judged to be passable. A computer program for causing a computer to function as a determination device comprising a determination unit that determines whether the quality of the produce subject to determination is pass or fail based on a determination model corresponding to the pass conditions determined by the pass condition determination unit and sensor information obtained for the produce subject to determination.

Images