Processed food product management device, model generation device, processed food product management method, model generation method, and program
A model-based approach using temperature and humidity variables calculates defect probability in kneaded product manufacturing, addressing defects by optimizing temperature settings for improved quality.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- 株式会社ニッスイ
- Filing Date
- 2020-06-09
- Publication Date
- 2026-07-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing methods for managing temperature conditions in the manufacturing of kneaded fishery products, such as fish cakes and fish paste, are prone to defects at the start of processing, necessitating improved methods to set appropriate temperature conditions.
A model is generated using temperature and humidity conditions as explanatory variables, along with pass/fail information, to calculate the probability of defective products occurring during the temperature control process, facilitating easier setting of optimal conditions.
Enables easier and more accurate setting of temperature conditions to reduce defects in kneaded products, ensuring quality consistency without requiring skilled personnel.
Smart Images
Figure 0007886680000001 
Figure 0007886680000002 
Figure 0007886680000003
Abstract
Description
Technical Field
[0004] ,
[0006] , , ,
[0005] , , , ,
[0001] The present invention relates to a kneaded product management device, a model generation device, a kneaded product management method, a model generation method, and a program.
Background Art
[0002] The manufacturing process of kneaded fishery products may include a temperature control process. For example, the kneaded product may be manufactured through a heating process after forming the minced fish as the raw material. Examples of the heating process include a baking process used for fish cakes, a frying process used for tempura, and a steaming process used for fish paste products. These heating processes may also include a sitting process to strengthen the binding between the fish meats. In such kneaded products, management of the conditions of the heating process is important. For example, Patent Document 1 describes comparing an image of a fish cake with a reference image and controlling the output of a fish cake baking machine using the comparison result.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0004] In the technique described in Patent Document 1, there remains a possibility that the kneaded product manufactured immediately after the start of processing becomes defective. In order to reduce such defects, it is necessary to set the temperature conditions to appropriate conditions even at the start of processing.
[0005] An object of the present invention is to make it easy to set the temperature conditions to appropriate values in the temperature control process of the kneaded product.
Means for Solving the Problems
[0006] According to the present invention, a storage unit stores a model that has been learned as ground truth data, in which temperature conditions, including at least one of the temperature and the value of the parameter controlling the temperature, and at least one of the humidity of the atmosphere in the temperature control process, are used as explanatory variables in a temperature control process included in the manufacturing process of processed seafood products, and pass / fail information indicating the quality of the processed seafood products manufactured is used as ground truth data. A calculation unit that acquires the temperature conditions and uses the temperature conditions and the model to calculate defect probability information indicating the probability of defective products occurring due to the temperature control process of the paste product under the acquired temperature conditions, A processed food management device equipped with the following is provided.
[0007] According to the present invention, a training data acquisition unit acquires multiple training data sets in which temperature conditions, including at least one of the temperature and the value of the parameter controlling the temperature in a temperature control process included in the manufacturing process of processed seafood products, and humidity in the atmosphere of the temperature control process, are used as explanatory variables, and pass / fail information indicating the quality of the manufactured processed seafood products is used as the correct answer data. A model generation unit generates a model for calculating the probability of defective products occurring due to the temperature control process of the processed product by learning the aforementioned multiple training data, Equipped with, The aforementioned model provides a model generation device that calculates the probability of occurrence using the temperature conditions.
[0008] According to the present invention, a computer, In the temperature control process included in the manufacturing process of processed seafood products, at least one of the baking temperature and the value of the parameter controlling the baking temperature, and at least one of the humidity of the atmosphere in the temperature control process are used as explanatory variables, and a model is trained to store pass / fail information indicating the quality of the manufactured processed seafood products as ground truth data. A method for managing processed products is provided, which involves acquiring the aforementioned temperature conditions, and using those temperature conditions and the model to calculate defect probability information indicating the probability of defective products occurring in the processed product due to the temperature control process under the acquired temperature conditions.
[0009] According to the present invention, a computer, In the temperature control process included in the manufacturing process of processed seafood products, at least one of the temperature and the value of the parameter controlling the temperature, and at least one of the humidity of the atmosphere in the temperature control process are used as explanatory variables. Multiple training data sets are obtained, with pass / fail information indicating the quality of the manufactured processed seafood products as the ground truth data. By learning from the aforementioned multiple training data, a model is generated to calculate the probability of defective products occurring in the processed product due to the temperature control process. The aforementioned model provides a model generation method for calculating the probability of occurrence using the temperature conditions.
[0010] According to the present invention, a computer can, A memory function that stores a model learned as ground truth data, in which temperature conditions, including at least one of the temperature and the value of the temperature control parameter, and at least one of the humidity of the atmosphere in the temperature control process, are used as explanatory variables in a temperature control process included in the manufacturing process of processed seafood products, and pass / fail information indicating the quality of the manufactured processed seafood products is used as ground truth data. A calculation function that acquires the aforementioned temperature conditions and uses the aforementioned temperature conditions and the aforementioned model to calculate defect probability information indicating the probability of defective products occurring due to the temperature control process of the paste product under the acquired temperature conditions, A program is provided to give it this feature.
[0011] According to the present invention, a computer can, A training data acquisition function that acquires multiple training data sets, in which temperature conditions, including at least one of the temperature and the value of the parameter controlling the temperature, and at least one of the humidity of the atmosphere in the temperature control process, are used as explanatory variables, and pass / fail information indicating the quality of the manufactured processed seafood products is used as the ground truth data. A model generation function that generates a model for calculating the probability of defective products caused by the temperature control process of the kneaded product by learning the plurality of training data; be provided with The model is provided with a program for calculating the occurrence probability using the temperature conditions.
Advantages of the Invention
[0012] According to the present invention, even without a skilled person, it becomes easier to find temperature conditions for achieving a color tone and shape that meet the standards.
Brief Description of the Drawings
[0013] [Figure 1] It is a diagram for explaining the usage environment of the model generation device and the kneaded product management device according to the embodiment. [Figure 2] It is a diagram showing an example of a manufacturing device. [Figure 3] It is a diagram showing a specific example of a manufacturing device. [Figure 4] It is a diagram showing a part of the data stored in the data storage unit. [Figure 5] It is a diagram showing another part of the data stored in the data storage unit. [Figure 6] It is a diagram showing an example of the functional configuration of the model generation device. [Figure 7] It is a diagram showing a first example of the training data stored in the training data storage unit. [Figure 8] It is a diagram showing a second example of the training data stored in the training data storage unit. [Figure 9] It is a diagram showing an example of the functional configuration of the kneaded product management device. [Figure 10] It is a diagram showing an example of the hardware configuration of the model generation device. [Figure 11] It is a flowchart showing an example of the process performed by the model generation device. [Figure 12] It is a flowchart showing an example of the process performed by the kneaded product management device. [Modes for carrying out the invention]
[0014] Embodiments of the present invention will be described below with reference to the drawings. In all drawings, similar components are denoted by the same reference numerals, and their descriptions are omitted as appropriate.
[0015] Figure 1 is a diagram illustrating the operating environment of the model generation device 20 and the processed fish product management device 30 according to the embodiment. The model generation device 20 and the processed fish product management device 30 are used together with the processed fish product manufacturing device 40. The processing process for processed fish products by the manufacturing device 40 includes a temperature control step, for example, a heating step. In the following description, the heating step will be used as an example. The processed fish product management device 30 uses the conditions of this temperature control step (hereinafter referred to as temperature conditions; in the case of the heating step, it will be referred to as heating conditions) to calculate information indicating the probability of defective processed fish products occurring due to the heating step under those heating conditions (hereinafter referred to as defect probability information). In calculating the defect probability information, the processed fish product management device 30 uses a model. This model is generated by the model generation device 20. Examples of processed fish products include chikuwa (fish cake), but are not limited to these.
[0016] The model generation device 20 generates a model using training data. This model may be generated for each production line of the processed food product. Here, production devices having the same structure may be treated as the same production device, or they may be treated as different production devices.
[0017] In the example shown in this figure, the model generation device 20 and the processed food management device 30 are used together with the data storage unit 10. The data storage unit 10 stores various types of information obtained from the processed food production line. The information stored in the data storage unit 10 includes heating conditions and information indicating whether the processed food product meets or fails the quality requirements under those heating conditions (hereinafter referred to as pass / fail information).
[0018] Furthermore, the model generation device 20 and the processed product management device 30 may be a single device. In addition, the data storage unit 10, the model generation device 20, and the processed product management device 30 may be a single device.
[0019] Figure 2 is a schematic diagram of the configuration of the manufacturing apparatus 40. In the example shown in this figure, the manufacturing apparatus 40 has a molding unit 410, a heating unit 420, and an imaging unit 430. The molding unit 410 molds the fish paste containing minced seafood into a predetermined shape. The molded fish paste is transported to the heating unit 420. In the heating unit 420, the molded fish paste is subjected to a heat treatment. The imaging unit 430 images the fish paste during or after the heat treatment. The imaging unit 430 may generate both images of the fish paste during the heat treatment and images of the fish paste after the heat treatment.
[0020] The image generated by the imaging unit 430 is processed by the image processing unit 432. The image may be taken against a background with a single color to make it easier to judge the color changes that occur during the heating process, such as the browning and charring, or against a black or grayish-brown background to make it easier to judge the white of the fish paste. The image may be taken while illuminating the object from the front to prevent uneven coloring. When taking images while illuminating, warm white, white, or daylight white light may be used to prevent blurring in color judgment.
[0021] The image processing unit 432 processes the image and determines whether the color change that occurs during the heating process of the processed fish product meets the standard, that is, whether the quality of the processed fish product after the heating process meets the standard. The color change that occurs during the heating process of the processed fish product may include not only the overall color change during the heating process, but also, for example, the number and area of burnt parts per unit area during the baking process, the darkest color of the burnt parts during the baking or frying process, and combinations thereof, which may be used as quality standards. Furthermore, the image processing unit 432 may also determine whether the shape of the processed fish product after the heating process meets the standard. This determination result, i.e., the pass / fail information described above, is stored in the data storage unit 10 along with the image. Note that the image processing unit 432 may be provided in the model generation device 20.
[0022] Figure 3 shows a specific example of the manufacturing apparatus 40. The manufacturing apparatus 40 shown in this figure is a chikuwa (fish cake) manufacturing apparatus. In the example shown in this figure, the heating process section 420 is located in the middle of the conveyor line. In the heating process section 420, the conveyor line repeatedly bends 180° from top to bottom, forming multiple stages 422. A heater 424 is provided below the lowest stage 422. Therefore, the quality of the processed product due to the heating process (for example, the browning after the baking process) is easily affected by the temperature of the lowest stage 422. The manufacturing apparatus 40 has a thermometer to measure the temperature of at least the lowest stage 422. The manufacturing apparatus 40 may have a thermometer for each of the stages 422. The manufacturing apparatus 40 may have a thermometer at the beginning of the heating process section 420. The manufacturing apparatus 40 may also have a thermometer at the end of the heating process section 420. In this case, the measurement results of each thermometer are stored in the data storage unit 10. The measurement results from these thermometers represent at least a part of the heating conditions and are stored in the data storage unit 10.
[0023] Furthermore, if the settling process of the processed product does not require physical heating and is a cooling or room temperature process, the manufacturing apparatus 40 may measure the temperature of the settling process and set it as part of the temperature control process. The settling process of the processed product affects the hardness of the processed product and can affect the quality caused by the temperature control process. Measuring the temperature of the settling process as part of the temperature control process makes it easier to find more appropriate temperature control conditions.
[0024] In addition to the thermometer described above, sensors for measuring other heating conditions are provided in or around the manufacturing apparatus 40. An example of such a sensor is a hygrometer. This hygrometer measures the humidity in the atmosphere of the heating process section 420. The hygrometer may also measure the humidity of the space in which the manufacturing apparatus 40 is installed. As sensors for measuring heating conditions, at least one of the following may be provided: a pressure gauge, an illuminometer, an odor meter, an oxygen concentration meter, a carbon dioxide meter, and a nitrogen concentration meter.
[0025] A scanning unit 430 is provided downstream of the heating process unit 420. The scanning unit 430 takes images of the processed product after it has undergone the heating process.
[0026] Figure 4 shows a portion of the data stored in the data storage unit 10. As shown in this figure, the data storage unit 10 stores images generated by the imaging unit 430 of the manufacturing apparatus 40, along with information indicating the date and time the image was generated, and pass / fail information for the processed products depicted in the image. Note that if the image processing unit 432 is provided in the model generation apparatus 20, the data storage unit 10 may not store the pass / fail information.
[0027] Figure 5 shows another portion of the data stored in the data storage unit 10. As shown in this figure, the data storage unit 10 stores data generated by the thermometer and hygrometer of the manufacturing apparatus 40, associating it with the date and time the data was generated. Furthermore, if the manufacturing apparatus 40 is equipped with sensors other than the thermometer and hygrometer, the data storage unit 10 also stores data generated by these sensors, associating it with the date and time the data was generated.
[0028] The data storage unit 10 may also store parameter values for controlling the heating temperature of the processed food product, either along with or in place of the data generated by the thermometer. An example of such parameter is the control data for the heater 424 (e.g., target temperature or heater output). In this case, the model generation device 20 and the processed food product management device 30 use the control data for the heater 424, either along with or in place of the data generated by the thermometer.
[0029] Figure 6 shows an example of the functional configuration of the model generation device 20. The model generation device 20 comprises a training data storage unit 210, a training data acquisition unit 220, and a model generation unit 230.
[0030] The training data storage unit 210 stores training data for generating the model. The training data uses the heating conditions in the heating process section 420 of the manufacturing apparatus 40 as explanatory variables, and the pass / fail information of the manufactured processed products as ground truth data. The training data is generated using data stored in the data storage unit 10.
[0031] As described above, in the data storage unit 10, the heating conditions are linked to the date and time when the heating conditions were measured. Similarly, in the data storage unit 10, the pass / fail information is also linked to the date and time when the image from which the pass / fail information originated was generated. Therefore, by using the date and time stored in the data storage unit 10, it is possible to link the pass / fail information with the corresponding heating conditions (i.e., the same date and time) and generate training data. The generation of training data is performed, for example, by the training data acquisition unit 220.
[0032] Furthermore, the training data acquisition unit 220 acquires training data from the training data storage unit 210 at the time of model generation. The model generation unit 230 generates a model by learning from the training data acquired by the training data acquisition unit 220. This model calculates the probability of defective products occurring due to the heating process of the processed food product using heating conditions.
[0033] In the process described above, the heating conditions included in the training data include at least one of the temperature of the heating process section 420 and the parameters that control this temperature, and at least one of the atmosphere of the heating process section 420 (or the space in which the manufacturing apparatus 40 is installed), but it is preferable that both of these be included.
[0034] Furthermore, the training data acquisition unit 220 may use only the heating conditions when the pass / fail information indicates failure as training data.
[0035] Figure 7 shows a first example of training data stored in the training data storage unit 210. In the example shown in this figure, the training data is stored in association with explanatory variables (including at least heating conditions) and pass / fail information for those explanatory variables.
[0036] Figure 8 shows a second example of training data stored in the training data storage unit 210. In this example, the training data consists only of explanatory variables for when the pass / fail information is "fail".
[0037] Figure 9 shows an example of the functional configuration of the processed food management device 30. In the example shown in this figure, the processed food management device 30 includes a memory processing unit 310, a model storage unit 320, a calculation unit 330, and a display 340. The memory processing unit 310 acquires the model generated by the data storage unit 10 and stores it in the model storage unit 320. The calculation unit 330 acquires the current or future heating conditions of the heating process unit 420 of the manufacturing apparatus 40 and generates defect probability information under those heating conditions by inputting the heating conditions into the model. As described above, the defect probability information is information that indicates the probability of defective products occurring due to the heating process.
[0038] The heating conditions used by the calculation unit 330 are input by, for example, the user of the processed food management device 30, but the current or future settings of the heating process unit 420 may also be read.
[0039] Figure 10 shows an example of the hardware configuration of the model generation device 20. The model generation device 20 includes a bus 1010, a processor 1020, a memory 1030, a storage device 1040, an input / output interface 1050, and a network interface 1060.
[0040] Bus 1010 is a data transmission path for the processor 1020, memory 1030, storage device 1040, input / output interface 1050, and network interface 1060 to send and receive data to and from each other. However, the method of connecting the processor 1020 and the other components to each other is not limited to bus connection.
[0041] Processor 1020 is a processor implemented in components such as the CPU (Central Processing Unit) and GPU (Graphics Processing Unit).
[0042] Memory 1030 is a main memory device implemented using RAM (Random Access Memory), etc.
[0043] The storage device 1040 is an auxiliary storage device implemented as an HDD (Hard Disk Drive), SSD (Solid State Drive), memory card, or ROM (Read Only Memory). The storage device 1040 stores program modules that implement each function of the model generation device 20 (for example, the training data acquisition unit 220 and the model generation unit 230). The processor 1020 reads these program modules into the memory 1030 and executes them, thereby realizing each function corresponding to that program module. The storage device 1040 also functions as a training data storage unit 210.
[0044] The input / output interface 1050 is an interface for connecting the model generation device 20 with various input / output devices. For example, the model generation device 20 communicates with the processed product management device 30 via the input / output interface 1050.
[0045] The network interface 1060 is an interface for connecting the model generation device 20 to a network. This network may be, for example, a LAN (Local Area Network) or a WAN (Wide Area Network). The network interface 1060 may connect to the network via a wireless connection or a wired connection. The model generation device 20 may communicate with the processed product management device 30 via the network interface 1060.
[0046] The hardware configuration example of the processed food management device 30 is the same as the example shown in Figure 9. In this case, the storage device 1040 stores program modules that realize each function of the processed food management device 30 (for example, the memory processing unit 310 and the calculation unit 330). The storage device 1040 can also function as a model storage unit 320.
[0047] Figure 11 is a flowchart showing an example of the processing performed by the model generation device 20. When models are generated for each manufacturing device 40 (or for each type of manufacturing device 40), the model generation device 20 performs the processing shown in this figure for each manufacturing device 40 (or for each type of manufacturing device 40).
[0048] First, the training data acquisition unit 220 of the model generation device 20 reads training data from the training data storage unit 210 (step S10). Prior to step S10, the training data acquisition unit 220 periodically performs the training data generation process. However, the training data acquisition unit 220 may also perform the training data generation process in step S10. In this case, the training data generation process becomes the training data acquisition process.
[0049] Next, the model generation unit 230 generates a model using the training data acquired by the training data acquisition unit 220 (step S20). Then, the model generation unit 230 outputs the generated model to the paste product management device 30 and stores it in the model storage unit 320 of the paste product management device 30 (step S30).
[0050] Figure 12 is a flowchart illustrating an example of the processing performed by the processed food management device 30. First, the calculation unit 330 of the processed food management device 30 reads a model from the model storage unit 320. Next, the calculation unit 330 obtains current or future heating conditions and inputs these heating conditions into the model to calculate defect probability information (step S120). Then, the calculation unit 330 displays the calculated defect probability information on the display 340 (step S130).
[0051] The user of the processed food management device 30 determines whether the defect probability information displayed on the display 340 is within an acceptable range. If it is outside the acceptable range (step S140: No), the user returns to step S110. In step S110, the modified heating conditions are entered. Note that the determination shown in step S140 may also be performed by the calculation unit 330.
[0052] Furthermore, the calculation unit 330 of the processed food management device 30 may control the heating process unit 420 of the manufacturing device 40 when it obtains the current heating conditions from the manufacturing device 40 and performs the processing shown in Figure 12. For example, after the calculation unit 330 becomes No in step S140 of Figure 12, if the defect probability information for the changed heating conditions falls within an acceptable range, it may output those heating conditions to the manufacturing device 40. In this case, after the heating process unit 420 of the manufacturing device 40 obtains the heating conditions from the processed food management device 30, it performs the heating process according to these heating conditions. Doing so can suppress a decrease in the yield of processed food products.
[0053] As described above, according to this embodiment, the model generation device 20 generates a model for calculating the probability of a processed food product being defective based on the heating conditions of the processed food product. The processed food product management device 30 then uses this model to calculate the probability of a processed food product being defective based on the current or future heating conditions. This makes it easier to set the heating conditions to appropriate values during the heating process of the processed food product.
[0054] The embodiments of the present invention have been described above with reference to the drawings, but these are merely examples of the present invention, and various other configurations can also be adopted.
[0055] Furthermore, while the flowcharts used in the above description show multiple steps (processes) in sequence, the execution order of the steps performed in each embodiment is not limited to the order in which they are described. In each embodiment, the order of the illustrated steps can be changed to the extent that it does not impede the content. Also, the above embodiments can be combined to the extent that their contents do not conflict. [Explanation of Symbols]
[0056] 10 Data storage unit 20 Model Generators 30 Processed food management device 40 Manufacturing equipment 410 Molding section 420 Heating process section 430 Imaging Unit 432 Image Processing Unit 210 Training data storage unit 220 Training Data Acquisition Unit 230 Model Generation Unit 310 Memory Processing Unit 320 Model Memory Unit 330 Calculation Unit 340 displays
Claims
1. A storage unit stores a model that has been trained as ground truth data, in which temperature conditions, including at least one of the temperature and the value of the parameter controlling the temperature, and the humidity of the atmosphere in the temperature control process, are used as explanatory variables, and pass / fail information indicating the quality of the manufactured processed seafood product is used as ground truth data. A calculation unit that acquires the temperature conditions and uses the temperature conditions and the model to calculate defect probability information indicating the probability of defective products occurring due to the temperature control process of the paste product under the acquired temperature conditions, Equipped with, A processed food management device in which the criteria for determining whether a product is acceptable or unacceptable in terms of quality and whether it is a defective product relate to the change in color that occurs during the temperature control process.
2. In the processed food management device according to claim 1, The aforementioned model is generated by learning the explanatory variables when the pass / fail information indicates failure, and is a processed food management device.
3. In the processed food management device according to either claim 1 or 2, In the temperature control step, the paste product moves along the line, The aforementioned line has multiple overlapping steps, The aforementioned processed food product was moved from the upper shelf to the lower shelf. The heat source is located below the lowest level. The aforementioned temperature is the temperature of the lowest stage in the processed food management device.
4. In the processed food management device according to any one of claims 1 to 3, The calculation unit described above, If the aforementioned failure probability information falls outside the acceptable range, the failure probability information is recalculated after the temperature conditions are changed. A processed food management device that, when the recalculated defect probability information falls within the acceptable range, outputs the modified temperature conditions to the device that performs the temperature control process.
5. A training data acquisition unit acquires multiple training data sets, in which temperature conditions, including at least one of the temperature and the value of the parameter controlling the temperature in the temperature control process included in the manufacturing process of processed seafood products, and the humidity of the atmosphere in the said temperature control process, are used as explanatory variables, and pass / fail information indicating the quality of the manufactured processed seafood products is used as the ground truth data. A model generation unit generates a model for calculating the probability of defective products occurring due to the temperature control process of the processed product by learning the aforementioned multiple training data, Equipped with, The aforementioned model calculates the probability of occurrence using the aforementioned temperature conditions, A model generation apparatus in which the criteria for determining whether a product is acceptable or unacceptable in terms of quality and whether it is a defective product relate to the change in color tone that occurs during the temperature control process.
6. In the model generation apparatus according to claim 5, The aforementioned pass / fail information is generated by a model generation device that processes images of the processed product after the temperature control process.
7. In the model generation apparatus according to claim 6, The pass / fail information is linked to the date and time when the image from which the pass / fail information originated was generated. The temperature conditions in the aforementioned temperature control process are linked to the date and time on which the temperature conditions were measured. The training data acquisition unit is a model generation device that generates the training data by linking the pass / fail information and the temperature conditions using the date and time.
8. In the model generation apparatus according to any one of claims 5 to 7, The aforementioned model is generated by a model generation device that learns the explanatory variables when the pass / fail information indicates failure.
9. In the model generation apparatus according to any one of claims 5 to 8, In the temperature control step, the paste product moves along the line, The aforementioned line has multiple overlapping steps, The aforementioned processed food product was moved from the upper shelf to the lower shelf. The heat source is located below the lowest level. The aforementioned temperature is the temperature of the lowest stage of the model generation device.
10. Computers In the temperature control process included in the manufacturing process of processed seafood products, at least one of the temperature and the value of the parameter controlling the temperature, and the humidity of the atmosphere in the temperature control process are used as explanatory variables, and a model is trained to store pass / fail information indicating the quality of the manufactured processed seafood products as ground truth data. The temperature conditions are obtained, and using the temperature conditions and the model, defect probability information is calculated that indicates the probability of defective products occurring due to the temperature control process of the paste product under the obtained temperature conditions. A method for managing processed food products, wherein the criteria for determining whether a product is acceptable or unacceptable in terms of quality and whether it is a defective product relate to changes in color that occur during the temperature control process.
11. Computers In the temperature control process included in the manufacturing process of processed seafood products, at least one of the temperature and the value of the parameter controlling the temperature, along with the humidity of the atmosphere in the temperature control process, are used as explanatory variables. Multiple training data sets are obtained, with pass / fail information indicating the quality of the manufactured processed seafood products as the ground truth data. By learning from the aforementioned multiple training data, a model is generated to calculate the probability of defective products occurring in the processed product due to the temperature control process. The aforementioned model calculates the probability of occurrence using the aforementioned temperature conditions, A model generation method in which the criteria for determining whether a product is of acceptable or unacceptable quality and whether it is a defective product relate to the change in color tone that occurs during the temperature control process.
12. On the computer, A memory function that stores a model learned as ground truth data, in which temperature conditions, including at least one of the temperature and the value of the parameter controlling the temperature, and the humidity of the atmosphere in the temperature control process, are used as explanatory variables in the temperature control process included in the manufacturing process of processed seafood products, and pass / fail information indicating the quality of the manufactured processed seafood products is stored as ground truth data. A calculation function that acquires the aforementioned temperature conditions and uses the aforementioned temperature conditions and the aforementioned model to calculate defect probability information indicating the probability of defective products occurring due to the temperature control process of the paste product under the acquired temperature conditions, Give it to him A program in which the criteria for determining whether a product is of acceptable or defective quality relate to the color change that occurs during the temperature control process.
13. On the computer, A training data acquisition function that acquires multiple training data sets, in which temperature conditions, including at least one of the temperature and the value of the temperature control parameter, and the humidity of the atmosphere in the temperature control process, are used as explanatory variables, and pass / fail information indicating the quality of the manufactured processed seafood product is used as the ground truth data. A model generation function that generates a model for calculating the probability of defective products occurring due to the temperature control process of the processed product by learning from the aforementioned multiple training data, Give it to him The aforementioned model calculates the probability of occurrence using the aforementioned temperature conditions, A program in which the criteria for determining whether a product is of acceptable or defective quality relate to the color change that occurs during the temperature control process.