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

Abstract

To provide an information processing device capable of more accurately determining an iron scrap group containing iron scrap of one or more grades.SOLUTION: One embodiment of the present invention is an information processing device for determining a grade ratio of each grade included in an iron scrap group to be determined, and includes: a receiving unit that receives a set of iron scrap group data obtained by imaging the iron scrap group to be determined along with the progress of unloading from a loading bed; and a grade determination unit that uses a learning model created by learning using a set of learning image data obtained by imaging the iron scrap group containing iron scrap of one or more grades along with the progress of unloading from the loading bed and a label including grade ratios of the one or more grades associated with the entire iron scrap group being a target of the set of learning data and determines a grade ratio of each grade included in the entire iron scrap group to be determined while using the set of iron scrap group data as an input to the learning model.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing method, and an information processing program.

Background Art

[0002] When producing iron from iron ore, a huge amount of carbon dioxide is emitted in the process of ironmaking using iron ore and coal in a blast furnace. At present, with the demand for carbon dioxide reduction, the recycling of iron scrap has become more important.

[0003] Here, the iron scrap collected by a recycler is accumulated at a collection yard and then transported to an electric furnace manufacturer or the like, and the price is calculated based on the measured weight, the grade of the iron scrap, and the like.

[0004] Traditionally, at electric furnace manufacturers' scrap receiving sites, the proportion of grades of incoming scrap iron was assessed visually by inspectors, which was extremely inefficient. A grade determination system has been proposed that, each time a portion of the scrap iron piled on the bed of a scrap loading vehicle is lifted by a lifting magnet, the remaining scrap iron on the bed and the scrap iron lifted by the lifting magnet are photographed. A partial proportion determination unit averages the grade proportion estimated from the image generated by a vehicle camera just before lifting and the grade proportion estimated from the image generated by a lifting magnet camera at the time of lifting to determine the grade proportion of the lifted scrap iron. A total proportion determination unit then aggregates the grade proportions of the scrap iron determined by the partial proportion determination unit each time the lifting magnet lifts scrap iron, calculates the average value, and determines the grade proportion of the entire scrap iron unloaded from the scrap loading vehicle to the scrap yard. A partial proportion determination unit estimates the grade proportion for each image representing the iron scrap lifted by the lifting magnet from different viewpoints, and determines the grade proportion for each lifted iron scrap as an average of these estimates. A total proportion determination unit then further averages these grade proportions for each lifted iron scrap to determine the grade proportion for the entire iron scrap (see Patent Document 1). [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2020-95709 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] However, there is room for improvement in the accuracy of conventional technology in identifying iron scrap.

[0007] The present invention has been made in view of these points, and its objective is to provide an information processing device, an information processing method, or an information processing program that can more accurately determine a group of iron scrap containing one or more grades of iron scrap. [Means for solving the problem]

[0008] To solve these problems, a first aspect of the present invention is a method for determining the grade ratio of each grade included in a group of scrap iron to be judged, comprising the steps of: receiving a set of scrap iron data, which is imaged along the progress of unloading the scrap iron to be judged from a loading platform; and using a learning model created by learning with a set of training image data, which is imaged along the progress of unloading a group of scrap iron including one or more grades of scrap iron, from a loading platform, and a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is a label, which is

[0009] Furthermore, a second aspect of the present invention is the method of the first aspect, wherein the determination step includes a step of obtaining correlation values ​​between features extracted from a plurality of image data contained in the input set of scrap iron data as intermediate features.

[0010] Furthermore, a third aspect of the present invention is the method of the first aspect, wherein the determination step includes a step of determining a grade that has been pre-associated with a predetermined steel material when the group of iron scrap to be determined includes a predetermined steel material.

[0011] Furthermore, a fourth aspect of the present invention is a method according to the first or second aspect, further comprising the step of determining the types of iron scrap included in the iron scrap group to be determined based on the set of iron scrap group data.

[0012] Furthermore, a fifth aspect of the present invention is the method of the first or second aspect, wherein the learning model is created using learning sound data obtained by recording the sounds produced when each grade of scrap iron is dropped, and the determination is further input to the learning model using sound data obtained by recording the sounds produced when the group of scrap iron to be determined is dropped.

[0013] Furthermore, a sixth aspect of the present invention is a method according to the first or second aspect, further comprising the step of detecting foreign matter contained in the group of iron scrap to be determined.

[0014] Furthermore, a seventh aspect of the present invention is the method of the sixth aspect, wherein the learning model is created using composite data obtained by combining an image of a foreign object with an image of at least one of the set of learning image data.

[0015] Furthermore, an eighth aspect of the present invention is a method according to the first or second aspect, further comprising the step of calculating the value of the group of iron scrap to be determined according to the proportion of each grade included in the group of iron scrap to be determined.

[0016] Furthermore, a ninth aspect of the present invention is a program for causing a computer to execute a method for determining the grade ratio of each grade included in a group of scrap iron to be judged, the method comprising the steps of: receiving a set of scrap iron data, which is imaged of the group of scrap iron to be judged as it progresses to be unloaded from a loading platform; and using a learning model created by learning with a set of learning image data, which is imaged of a group of scrap iron including one or more grades of scrap iron, as it progresses to be unloaded from a loading platform, and a label which includes the grade ratio of one or more grades associated with the entire group of scrap iron that is the subject of the set of learning data, the steps of using the set of scrap iron data as input to the learning model to determine the grade ratio of each grade included in the entire group of scrap iron to be judged.

[0017] Moreover, a tenth aspect of the present invention is an apparatus for determining the grade ratio of each grade included in an iron scrap group to be determined, which receives a set of iron scrap group data obtained by imaging the iron scrap group to be determined along the progress of unloading from the loading platform, and a set of learning image data obtained by imaging an iron scrap group including iron scraps of one or more grades along the progress of unloading from the loading platform, and uses a learning model created by learning using a label including the grade ratio of one or more grades associated with the entire iron scrap group that is the target of the set of learning data. The grade ratio of each grade included in the entire iron scrap group to be determined is determined by using the set of iron scrap group data as an input to the learning model.

Advantages of the Invention

[0018] According to one aspect of the present invention, it is possible to provide an information processing apparatus, an information processing method, or an information processing program capable of more accurately determining an iron scrap group.

Brief Description of the Drawings

[0019] [Figure 1] It is a schematic configuration diagram of an iron scrap yard according to an embodiment of the present invention. [Figure 2] It is a diagram showing an example of the schematic configuration of an information processing system according to this embodiment. [Figure 3] It is a diagram showing an example of the hardware configuration of a server according to this embodiment. [Figure 4] It is a diagram showing an example of the functional configuration of a server according to this embodiment. [Figure 5] It is a diagram showing an example of a display screen displayed on a display device of a user terminal according to this embodiment. [[ID=I29]] [Figure 6] It is a diagram showing an example of a display screen displayed on a display device of a user terminal according to this embodiment. [Figure 7] It is a flowchart showing an example of a process executed by a server according to this embodiment. [Figure 8] It is a graph showing the determination result by the learning model according to this embodiment. [Figure 9] It is a graph showing the determination result by the learning model according to the prior art as a comparative example.

Embodiments for Carrying Out the Invention

[0020] As shown in the schematic configuration diagram of FIG. 1, an iron scrap yard according to an embodiment of the present invention controls the magnetic force to adsorb and release it with respect to an iron scrap group 10 including one or more grades, which is loaded on the loading platform of a truck 40 parked at a predetermined position. It has a lift magnet 30 and a crane rail 50 for moving a part of the iron scrap group 10 adsorbed by the lift magnet 30 to the iron scrap placement yard 20. Further, the iron scrap yard may have a measuring device (not shown) for measuring the weight of the truck 40 loaded with the iron scrap group 10 on the loading platform. A value obtained by subtracting the weight of the truck 40 from the weight measured by this measuring device can be input to the server 3 shown in FIG. 2 as the weight of the iron scrap group 10.

[0021] The information processing system 1 according to the present embodiment has one or more imaging devices 2 for photographing the iron scrap group 10 loaded on the loading platform of the truck 40. As shown in FIG. 2, each imaging device 2 is connected to a server or an information processing device 3 via a network 5 so as to be capable of data communication. Further, the server 3 is connected to the user terminal 4 via the network 5 so as to be capable of data communication. The number of the server 3, the user terminal 4 and the imaging device 2 connected via the network 5 to this server 3 is arbitrary respectively.

[0022] (Imaging device 2) The imaging device 2 is positioned to preferably photograph the group of scrap iron 10 from directly above, and photographs the group of scrap iron 10, transmitting the photographed scrap iron data (image data) to the server 3. As an example, the imaging device 2 photographs the surface of the scrap iron 10 remaining on the bed of a truck 40 parked at a predetermined location each time the lift magnet 30 lifts some of the scrap iron. The imaging device 2 repeats this photography of newly appearing surfaces until all the scrap iron is removed from the bed of the truck 40. In this way, multiple sets of scrap iron data are obtained, photographed in accordance with the progress of the unloading of the scrap iron 10 from the bed by the lift magnet 30.

[0023] Furthermore, it is preferable that the imaging device 2 be installed in a position that looks directly down at the group of scrap iron 10 loaded on the bed of a truck 40 parked at a predetermined location. However, if a predetermined object in the scrap iron yard, such as a lift magnet 30 or a crane rail 50, is within the imaging range during imaging, the imaging device 2 may be installed in a position where the lift magnet 30 or crane rail 50 is outside the imaging range or its central part.

[0024] (Server 3) Figure 3 shows an example of the hardware and functional configuration of server 3. As shown in Figure 3, server 3 is equipped with a communication IF 300A, a storage device 300B, a CPU 300C, etc. Server 3 may also be equipped with input devices (e.g., keyboard, touch panel, etc.) and display devices (e.g., LCD monitor, OLED monitor, etc.).

[0025] The communication IF300A is an interface for communicating with an external terminal (in this embodiment, a user terminal 4, etc.).

[0026] The storage device 300B is, for example, an HDD or a semiconductor storage device. The storage device 300B stores information processing programs and various databases used by the server 3. For example, the storage device 300B stores table data for calculating the value of the iron scrap group 10. For example, the storage device 300B stores the unit price (price per predetermined weight) according to the grade of the iron scrap for each type. In this embodiment, the information processing programs and various databases are stored in the storage device 300B of the server 3, but they may also be stored in an external storage device such as a USB memory or an external server connected via a network, and configured to be accessible or downloadable as needed.

[0027] The CPU300C controls server 3 and includes, for example, ROM (Read Only Memory) and RAM (Random Access Memory), which are not shown in the diagram.

[0028] As shown in Figure 4, the server 3 has functions such as a receiving unit 301, a transmitting unit 302, a storage device control unit 303, a product type determination unit 304, a grade determination unit 305, a foreign object detection unit 306, and a value calculation unit 307. Each process or operation shown in Figure 4 is realized by the CPU 300C executing instructions contained in an information processing program stored in the server 3's storage device 300B, etc. This program may contain one or more programs and can be recorded on a computer-readable storage medium to become a non-transient program product.

[0029] The receiving unit 301 receives, for example, imaging data (image data) of the group of scrap iron 10 transmitted from the imaging device 2. The receiving unit 301 also receives, for example, inspection information from the supplier of the scrap iron 10 (see left side of Figure 5). Furthermore, the receiving unit 301 receives, for example, information on the weight of the scrap iron 10.

[0030] The transmitting unit 302 transmits information to the user terminal 4, such as the type (e.g., bales, heavy iron, shredder, new cuts, old pig iron, etc.) and grade determination results of the iron scrap group 10 on the server 3, the detection results of foreign matter contained in the iron scrap group 10, and the calculated value of the iron scrap group 10. The transmitting unit 302 can also transmit information that displays the detected foreign matter in a recognizable manner.

[0031] The memory control unit 303 controls the memory device 300B. For example, the memory control unit 303 writes and reads information from the memory device 300B.

[0032] The type determination unit 304 uses a learning model created by machine learning using, for example, training image data of a group of scrap iron and a label including the type associated with the training image data. The receiving unit 301 takes the scrap iron data as input to the learning model and can determine the type of scrap iron included in the scrap iron group 10 to be determined. The "label" is the type of scrap iron, for example, bales, heavy scrap, shredder, new cuts, or slag. The "label" may also be assigned a type determined by, for example, an inspector. Note that the type may not be determined by the type determination unit 304, but rather by a user, such as a worker, inputting it from a user terminal 4.

[0033] The grade determination unit 305 may use a learning model corresponding to the variety determined by the variety determination unit 304 or the variety entered by the user terminal 4 (for example, by the user), or it may use a single learning model instead of preparing a separate learning model for each variety.

[0034] The grade determination unit 305 may use, for example, a learning model created by machine learning using a set of training image data, which is an image of a group of scrap iron containing one or more grades of scrap iron, taken along with the progress of unloading from the loading platform, and a label corresponding to the set of training image data, which includes the grade ratio of one or more grades associated with the entire group of scrap iron. The receiving unit 301 can also use a set of scrap iron data, which is an image of the group of scrap iron to be determined, taken along with the progress of unloading from the loading platform, as input to the learning model to determine the proportion of each grade included in the entire group of scrap iron 10. By using the grade ratio associated with the entire group of scrap iron that is the subject of the set of training image data as a label for a set of training image data taken along with the progress of unloading from the loading platform, annotation of labels for each set of training image data becomes unnecessary. In this case, it is not necessarily required to use all of a set of training image data when learning for variety determination, foreign object detection, or other processing or operations, nor is it necessarily required to use all of a set of scrap iron data when learning for variety determination, foreign object detection, or other processing or operations.

[0035] While the conventional method, which involves determining the grade ratio of the iron scrap 10 identified by each training image data captured during the process of the lift magnet 30 unloading from the truck bed 40, and then using the result of this determination as annotation for each training image data to determine the overall grade ratio of the iron scrap 10, might seem appropriate for machine learning because it involves using a larger number of annotations, it is a task not present in the conventional visual inspection of each individual piece of scrap by the receiving officer. This results in a significant burden in creating numerous sets of training image data necessary for machine learning, and it is not a method that can practically improve accuracy through data accumulation. Furthermore, the inventors noted that there is a difference in the assessment accuracy of trainees regarding the proportion of each grade included in the iron scrap group between skilled and unskilled workers, and that high accuracy cannot be expected even from skilled workers in the new task of assessment for each training image data. Therefore, the reliability of annotations for each training image data is low, and consequently, the reliability of the learning model obtained using such annotations is also low. Based on these observations, the inventors discovered the present invention, which does not necessarily require annotation of labels for each training image data. As described later, the present invention achieves higher judgment accuracy than the prior art, and further improvements are expected. It objectively verifies or replaces the subjective assessment work performed by receiving personnel, and contributes to the recycling of iron scrap, which will become increasingly important in the future.

[0036] The number of training image data in a set is preferably between 3 and 15. If the number is less than 3, there may be insufficient information for grade determination, increasing the risk of decreased inference accuracy in the learning model. While exceeding 15 may increase analysis time, the analysis can still be performed.

[0037] The "label" includes, for example, the proportion of each grade of heavy scrap contained in the iron scrap group 10 as a grade ratio. The grades are indicated, for example, HS, H1, H2, H3, and H4. The "label" may be assigned a value determined, for example, by the inspector, or an actual measured value. The proportion is, for example, a weight ratio, but is not limited to this; it may also be an area ratio or a volume ratio.

[0038] The grade determination unit 305 may also use a learning model created by training with data that associates a predetermined steel material with its grade as training data, to determine the grade of the predetermined steel material included in the iron scrap group 10 based on the imaging data. In this case, the learning model, which has been trained on the predetermined steel material, detects the predetermined steel material and determines the grade by referring to information that has been pre-linked to the steel material and its grade. In a scrap yard, a predetermined steel material may have a fixed grade assigned to it, which is one of the important criteria for grade determination. Therefore, by learning not only the grade but also information on specific steel materials, it becomes possible to make more accurate determinations.

[0039] The foreign object detection unit 306 can detect foreign objects in the scrap iron group 10 by using, for example, training image data of a group of scrap iron containing foreign objects (e.g., wire harnesses and motors containing large amounts of copper, tin, lead, zinc, etc.) and a learning model created by machine learning that uses the type of foreign object (such as a motor) associated with the training image data as a label. The receiving unit 301 takes the scrap iron group data as input to the learning model. Here, the training image data can include not only image data of a group of scrap iron containing foreign objects, but also composite data obtained by combining an image of a foreign object and an image of the scrap iron group. In this way, although it is usually difficult to obtain images of various patterns containing foreign objects, images of various patterns containing foreign objects can be generated by synthesis and used as composite data for training, thereby improving the efficiency of the learning model.

[0040] The value calculation unit 307 calculates the value of, for example, the group of iron scrap 10. Here, the value calculation unit 307 calculates the value of the group of iron scrap 10 based on the determination results of the type determination unit 304 and the grade determination unit 305. Here, the value calculation unit 307 calculates the value of the group of iron scrap 10 according to the proportion or weight of non-iron materials (for example, foreign matter, dust, moisture, etc.) contained in the group of iron scrap 10. Specifically, the value calculation unit 307 subtracts the weight of non-iron materials (foreign matter, dust, moisture, etc.) contained in the group of iron scrap 10 from the weight of the group of iron scrap 10 ("weight subtraction"), and then calculates the value of the group of iron scrap 10 by multiplying the weight of the type determined by the type determination unit 304 (the type may be entered by the user) and the weight of each grade determined by the grade determination unit 305 by the unit price. The value calculation unit 307 may automatically determine the proportion or weight of materials other than iron scrap included in the iron scrap group 10 from the imaging data received by the receiving unit 301, or it may use the proportion or weight of materials other than iron scrap entered by the user from the user terminal 4.

[0041] (Display screen example) Figures 5 and 6 are shown below. Book This figure shows an example of a screen displayed on the display device of the user terminal 4 according to the embodiment. The screens displayed on the display device of the user terminal 4 will be described below with reference to Figures 5 and 6. In the following description, identical components are denoted by the same reference numerals, and redundant explanations are omitted.

[0042] Figure 5 shows an example of the inspection information input screen 51. As shown in Figure 5, on the left side of the screen, inspection information, such as the supplier's name, trading company name, type, and unloading location, can be entered into input boxes 52 to 55. Also, as shown in Figure 5, on the right side of the screen, an image 56 captured by the imaging device 2 (an image looking down at the group of scrap iron 10 loaded on the truck bed 40) is displayed. Furthermore, when the user operates the input device of the user terminal 4 and selects the inspection start button 57, the analysis of the scrap iron group (for example, grade determination, value calculation, etc.) begins.

[0043] Figure 6 shows an example of the display screen for the inspection results. As shown in Figure 6, the left side of the inspection results screen 71 displays information such as the inspection start time, inspection end time, inspection duration, supplier, responsible trading company, responsible inspector, unloading location, weight of the iron scrap group 10 (before weight deduction), and weight of the iron scrap group 10 (after weight deduction). As shown in Figure 6, the right side of the inspection results screen displays information such as the type, grade, percentage of each grade, unit price, price, reason for discount, and purchase price (value) of the iron scrap group 10 that was inspected.

[0044] (Value calculation process) Figure 7 is a flowchart showing an example of the value calculation process performed on Server 3. The value calculation process performed on Server 3 will be explained below with reference to Figure 7.

[0045] (Step S701) The receiving unit 301 of server 3 receives the proportion or weight of non-iron materials (e.g., dust, moisture, etc.) in the iron scrap group 10. Alternatively, the proportion or weight of dust, moisture, etc. in the iron scrap group 10 may be automatically determined from the image data received by the receiving unit 301.

[0046] (Step S702) The value calculation unit 307 calculates the value of the scrap iron group 10. Specifically, the value calculation unit 307 subtracts the weight of non-scrappy materials (foreign matter, dust, moisture, etc.) contained in the scrap iron group 10 from the weight of the scrap iron group 10 ("weight subtraction"). Next, the value calculation unit 307 refers to the price data of the corresponding type in the value calculation table data stored in the storage device 300B. Next, the value calculation unit 307 calculates the weight of each grade contained in the scrap iron group 10 based on the weight of the scrap iron group 10 after weight subtraction and the proportion of each grade contained in the scrap iron group 10 as determined by the grade determination unit 305. Next, the value calculation unit 307 refers to the price data and calculates a total value for each grade by adding the weight multiplied by the unit price.

[0047] (Step S703) The transmission unit 302 transmits the calculated value of the scrap iron group 10 to the user terminal 4. As a result, the value of the scrap iron group is displayed on the display device of the user terminal 4, as illustrated in Figure 6.

[0048] Comparative experiment The results of assessments by inspectors for each truck, determining the proportion of each grade in the total amount of scrap iron loaded onto the truck bed, were prepared for 530 trucks. Additionally, the results of assessments performed by inspectors on five image data points taken by inspectors for each truck, capturing the surface of the scrap iron as it was unloaded from the truck bed, were prepared for 100 trucks. Furthermore, six grades, HS, H1, H2, H3, L1, and L2, were designated as judgment candidates. Using the former data, a learning model according to this embodiment (hereinafter also referred to as the "multiple input model") was generated, and using the latter data, a learning model according to the prior art (hereinafter also referred to as the "single input model") was generated. More specifically, in this embodiment, a model was constructed to predict the grade of a single truck by using ResNet50, a convolutional neural network, as the backbone and learning by using the correlation values ​​between features extracted from multiple image data as intermediate features. In the conventional technique, a model was constructed to predict the grade from a single image data, similarly using ResNet50 as the backbone, and the prediction of the grade of a single truck was calculated by averaging the predictions for multiple image data. In both methods, data partitioning was performed using k-fold cross-validation (k=5).

[0049] The evaluation method involved using the assessment results of each truck by the inspector as the answer (grand trueth (GT)), and evaluating it as correct (true) if the maximum error for each grade was within 10 points. Table 1 shows an example of the evaluation.

[0050] [Table 1]

[0051] Figure 8 shows the results for the multiple-input model, and Figure 9 shows the results for the single-input model. The former achieved an accuracy of 57.7%, while the latter achieved an accuracy of 45.0%, indicating that the former was approximately 12.7% more accurate.

[0052] In addition to the above embodiments, the type or grade of the iron scrap group 10 may be determined by a learning model created by learning using sound data generated when each type or grade of iron scrap is dropped. When using sound data generated when iron scrap is dropped, it is preferable to create learning sound data for each iron scrap yard.

[0053] If the learning model is created by further learning using sound data generated when iron scrap is dropped, a microphone is provided to capture the sound generated when the iron scrap group 10 is dropped into the iron scrap yard, and the server 3 receives the image data captured by the imaging device 2. to In addition, sound data is received that captures the sound produced when the group of iron scrap 10 is dropped.

[0054] Furthermore, in the embodiments described above, unless the word "only" is used, such as "based only on XX," "according only to XX," or "in the case of XX only," it is assumed in this specification that additional information may also be considered. Also, as an example, the statement "if a, then b" does not necessarily mean "always b in the case of a" or "b immediately after a," unless explicitly stated otherwise. Furthermore, the statement "each a constituting A" does not necessarily mean that A is composed of multiple components, but rather includes the possibility that the component is singular.

[0055] Furthermore, for the sake of clarity, even if there are aspects of operation in some method, program, terminal, device, server, or system (hereinafter referred to as "method, etc.") that differ from the operation described herein, each aspect of the present invention is intended to cover the same operation as any of the operations described herein, and the existence of operation different from the operation described herein does not mean that such method, etc. is outside the scope of each aspect of the present invention.

[0056] The flowchart shown in Figure 7 is merely an example and does not necessarily mean that the process will always start or end in the illustrated steps.

[0057] Furthermore, the above embodiments and modifications are merely examples of how the present invention may be implemented, and the technical scope of the invention should not be interpreted as being limited by them. In other words, the present invention can be implemented in various forms without departing from its gist or its main features. [Explanation of Symbols]

[0058] 1. Information Processing System 2. Imaging device 3 servers 4. User terminals 5 Network 10 Iron Scrap Group 20 Iron scrap storage area 30 Lift Magnets 40 tracks 50 Crane Rails 51 Information Input Screen 52, 53, 54, 55 Input boxes 56 Captured images 57 Start button 71 Analysis result screen 300A communication IF 300B storage device 300C CPU 301 Receiver 302 Transmitter 303 Storage Unit 304 Variety Judgment Department 305 Grade Judgment Department 306 Foreign object detection unit 307 Value Calculation Unit

Claims

1. A method for determining the grade ratio of each grade contained in a group of steel scrap to be evaluated, comprising the steps of: receiving a set of data on the group of iron scraps, the data being obtained by capturing a plurality of images of the group of iron scraps to be determined along with the progress of unloading from a loading platform; a step of using a learning model created by learning using a set of learning image data obtained by capturing multiple images of a group of iron scraps including one or more grades of iron scrap along the progress of unloading from a loading platform and labels including grade ratios of one or more grades associated with the entire group of iron scraps that is the subject of the set of learning data, and determining the grade ratios of each grade included in the entire group of iron scraps that is the subject of the determination by using the set of iron scrap group data as an input to the learning model; Includes.

2. 2. The method according to claim 1, wherein the determining step includes the step of obtaining, as an intermediate feature, a correlation value between features extracted from a plurality of image data included in the set of input iron scrap group data.

3. 2. The information processing device according to claim 1, The method further includes a step of determining a type of the iron scrap group to be determined based on the set of iron scrap group data, The learning model is a model corresponding to the determined variety.

4. 2. The method of claim 1 , The method further includes a step of detecting foreign matter contained in the group of iron scraps to be judged.

5. 5. The method of claim 4, The learning model is created using composite data obtained by combining an image of a foreign object with an image based on at least one of the set of learning image data.

6. 5. A method according to any one of claims 1 to 4, comprising: The method further includes a step of calculating a value of the group of iron scraps to be evaluated based on the proportion or weight of each grade of iron scrap contained in the group of iron scraps to be evaluated and the proportion or weight of materials other than the iron scrap contained in the group of iron scraps to be evaluated.

7. 5. A method according to any one of claims 1 to 4, comprising: The number of sets of learning image data is three or more.

8. A program for causing a computer to execute a method for determining a grade ratio of each grade contained in a group of iron scrap to be determined, the method comprising: receiving a set of data on the group of iron scraps, the data being obtained by capturing a plurality of images of the group of iron scraps to be determined along with the progress of unloading from a loading platform; a step of using a learning model created by learning using a set of learning image data obtained by capturing multiple images of a group of iron scraps including one or more grades of iron scrap along the progress of unloading from a loading platform and labels including grade ratios of one or more grades associated with the entire group of iron scraps that is the subject of the set of learning data, and determining the grade ratios of each grade included in the entire group of iron scraps that is the subject of the determination by using the set of iron scrap group data as an input to the learning model; Includes.

9. An apparatus for determining the grade ratio of each grade contained in a group of iron scrap to be evaluated, receiving a set of data on the group of iron scraps, the data being obtained by taking a plurality of images of the group of iron scraps to be judged along the progress of unloading from a loading platform; The present invention is configured to use a learning model created by learning using a set of learning image data, which is multiple images of a group of iron scrap containing one or more grades of iron scrap as it is unloaded from a loading platform, and labels including grade proportions of one or more grades associated with the entire group of iron scrap that is the subject of the set of learning data, and to determine the grade proportions of each grade contained in the entire group of iron scrap that is the subject of the determination by using the set of iron scrap group data as input to the learning model.

Images