Body weight measuring system and body weight measuring method

Abstract

To provide a body weight measuring system capable of measuring a body weight of an object animal appropriately.SOLUTION: A body weight measuring system which images an object animal with a camera 26 includes a machine learning model 41 with the plane dimension and width information 33 on the object animal and the side surface dimension and width information 34 on the object animal as explanatory variables, and with weight information on the object animal as objective variables, and a calibration part 60 for performing calibration of the machine learning model 41.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a weight measurement system and a weight measurement method. [Background technology]

[0002] The pig weight measuring device described in Patent Document 1 consists of a dustproof box attached to the ceiling of a pigpen near the water supply, a projector with a checkered slide attached, and a video camera installed inside the dustproof box. When light is projected from the projector, a checkered pattern is projected onto the floor and the surface of the pig's body. This image is captured by a video camera, and the pig's withers height (H) is calculated based on the misalignment of the checkered pattern in the captured image. The captured image is also binarized to black and white to determine the pig's projected area (A). There is a close relationship between the pig's withers height (H), projected area (A), and weight (W), expressed as W = aHbAc. Furthermore, the dimension of this multiple regression equation (d = b + 2c) is close to the dimension of volume (d = 3). Therefore, weight (W) is calculated using the withers height (H) and projected area (A).

[0003] Patent Document 2 discloses that a measuring device is provided that sets three selection criteria for fish during the fish farming process or at the time of shipping: body length only, body length and weight, and weight only, and that by selecting one of these criteria, the body length and weight of the fish can be measured efficiently and accurately without weakening them while they are still alive, allowing for multi-stage selection.

[0004] This measuring device is equipped with a measuring chute that slopes downward to allow a constant flow of water, and a passage detector attached midway to detect fish passing through the chute. Light sources and cameras capable of taking still images of fish based on the detector's detection signal are installed above and below the measuring chute. Meanwhile, a load cell that emits an output proportional to the fish's weight has its movable part attached to the measuring chute and its fixed part attached to the weighing frame. Images are processed using an image processing device and image memory built into the control device, and the image and load cell output are calculated into body length and weight data by the control device's CPU. The control device's sorting criteria setting device compares the input with three sorting criteria: body length only, body length and weight, and weight only, and a sorting device attached downstream sorts the fish into multiple stages.

[0005] Patent Document 3 describes the provision of an animal weight measuring device, a measuring facility, and an automatic measuring device.

[0006] This weighing device is composed of a load sensor, a weighing platform attached horizontally to the top of the sensor and of a predetermined size depending on the target animal, side fences arranged approximately vertically on the left and right ends of the weighing platform and of a predetermined height and depth depending on the target animal, and a front fence installed on the front end of the weighing platform and shorter than the side fences.The rear end between the side fences is open.The weighing facility is composed of one or more weighing devices arranged in a ring or horizontal row, a feeder 11 placed near the front of the front fence, and barrier members placed outside the side fences to prevent access to the feeder.The automatic weighing device is composed of a weighing device, a cable for transmitting the sensor signal, and a data collection and processing device. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Application Laid-Open No. 2002-286421 [Patent Document 2] Japanese Patent Application Publication No. 8-050052 [Patent Document 3] Japanese Patent Application Laid-Open No. 2003-114145 Summary of the Invention [Problem to be solved by the invention]

[0008] In the case of Patent Document 1, it is difficult to project the grid pattern onto the floor and the surface of the pig's body, and therefore appropriate measurements cannot always be made. The case of Patent Document 2 is based on measuring fish, and has the problem of being difficult to apply to measuring the weight of other animals.

[0009] The system described in Patent Document 3 requires the animals to be guided to a weighing platform, which is difficult. The biggest challenge facing pig farms is the aging population and declining working population. Weighing pigs requires a great deal of effort, which means that pigs cannot be weighed and are not shipped at the optimal time.

[0010] Furthermore, as shown in Figure 10(A), even when pigs are shipped at the same age, there are individual differences in growth even within the same pig farm, resulting in variations in weight as shown in Figure 10(B). The national average variation in meat grade after shipping is as shown in Figure 11. The area with high unit prices, enclosed by the frame W in Figure 11, is just over 50%, and the issue is how to increase this area.

[0011] For example, as shown in Figure 12(A), weight statistics for pigs raised on a certain farm and shipped on a certain day show that the number of pigs with a "high-grade" weight (70 to 80 kg) is less than 50%, missing the best point. If we could predict the optimal shipping date and ensure that the number of "high-grade" pigs exceeds 65%, as shown in Figure 12(B), we could potentially earn 450 yen per pig, which is the difference in unit price between Figure 12(A) and Figure 12(B). Furthermore, if we could predict the date and time when the number of "high-grade" pigs will exceed 65%, as shown in Figure 12(B), we could reduce feed costs. For example, reducing feed costs for eight days could save 1,600 yen per pig.

[0012] The embodiments of the present invention have been made in consideration of the above points, and have an object to provide a weight measurement system that can appropriately measure the weight of a target animal. [Means for solving the problem]

[0013] A weight measurement system according to an embodiment of the present invention comprises an image acquisition means for capturing an image of a target animal using a camera and obtaining a planar image from above and a side image from the side of the target animal; a planar dimension and area information acquisition means for obtaining planar dimension and area information of the target animal based on the planar image; a side dimension and area information acquisition means for obtaining side dimension and area information of the target animal based on the side image; a machine learning model obtained by machine learning using the planar dimension and area information and the side dimension and area information as explanatory variables and weight information of the target animal as a target variable, and a weight information acquisition means for providing the obtained planar dimension and area information and the side dimension and area information as explanatory variables to the machine learning model to obtain weight information, which is the target variable; and a calibration unit for calibrating the machine learning model.

[0014] A weight measurement method according to an embodiment of the present invention includes an image acquisition step of capturing an image of a target animal using a camera to obtain a planar image from above and a side image from the side of the target animal, a planar dimension and width information acquisition step of obtaining planar dimension and width information of the target animal based on the planar image, a side dimension and width information acquisition step of obtaining side dimension and width information of the target animal based on the side image, and a machine learning model obtained by machine learning using the planar width information, the planar length information, and the side dimension and width information as explanatory variables and weight information of the target animal as a target variable, and The method is characterized by comprising: a weight information acquisition step of providing the weight information to the machine learning model to obtain weight information as a dependent variable; a history information accumulation step of accumulating pairs of explanatory variables and dependent variables when weight information is obtained using the machine learning model in the weight information acquisition step as history information; and an updated learning model generation step of performing the same machine learning as when the machine learning model was created, based on input of actual weight information of the target animal corresponding to the explanatory variables in the pair in the history information accumulated in the history information accumulation step, using the actual weight information as a dependent variable and as an explanatory variable corresponding to this actual weight information, to obtain an updated machine learning model. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a configuration diagram of a weight measurement system according to an embodiment of the present invention configured using a computer. [Figure 2] FIG. 1 is a functional block diagram of a weight measurement system according to a first embodiment of the present invention. [Figure 3] FIG. 2 is an explanatory diagram of planar dimension / area information acquired by the weight measurement system according to the embodiment of the present invention. [Figure 4] FIG. 3 is an explanatory diagram of side dimension and area information acquired by the weight measurement system according to the embodiment of the present invention. [Figure 5] 3 is a flowchart showing the operation of the weight measurement system according to the first embodiment of the present invention. [Figure 6] FIG. 10 is a functional block diagram of a weight measurement system according to a second embodiment of the present invention. [Figure 7] FIG. 10 is an explanatory diagram of appropriate image selection by the weight measurement system according to the embodiment of the present invention. [Figure 8] FIG. 4 is an explanatory diagram of a correction process performed by the weight measurement system according to the embodiment of the present invention. [Figure 9] 10 is a flowchart showing the operation of a weight measurement system according to a second embodiment of the present invention. [Figure 10] FIG. 1 is a diagram showing the variation in weight due to growth of pigs, which are target animals in an embodiment of the present invention. [Figure 11] FIG. 1 is a diagram showing the variation in weight of dressed pigs, which are target animals in an embodiment of the present invention. [Figure 12] FIG. 1 is a diagram showing the variation in weight of pigs, which are target animals in an embodiment of the present invention, at the time of shipping, and the ideal weight distribution. [Figure 13] FIG. 10 is a diagram showing the estimation accuracy at a certain time of the weight measurement system according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, a weight measurement system and a weight measurement method according to an embodiment of the present invention will be described with reference to the accompanying drawings. In each drawing, the same components are designated by the same reference numerals, and duplicate explanations will be omitted. FIG. 1 shows a configuration diagram of a computer in which the weight measurement system according to an embodiment of the present invention is configured by a computer. That is, a CPU 10 configures the weight measurement system using programs and data in a main memory 11. An external memory interface 13, an input interface 14, a display interface 15, and a data input interface 16 are connected to the CPU 10 via a bus 12.

[0017] An external storage device 23 is connected to the external storage interface 13. The external storage device 23 stores programs and data for the operation of this weight measurement system, which can be read out and used by the CPU 10 in the main memory 11 as appropriate. For this reason, the external storage device 23 stores programs for implementing the machine learning models and various means described below. An input device 24 such as a keyboard or touch panel and a pointing device 22 such as a mouse are connected to the input interface 14. A display device 25 having a screen such as an LCD is connected to the display interface 15. Cameras 26-1 to 26-m are connected to the data input interface 16, and these cameras 26-1 to 26-m function to capture images of the target animal. The number of cameras 26-1 to 26-m is arbitrary. Image data obtained by the cameras 26-1 to 26-m is input by the data input interface 16 to required parts such as the CPU 10 and the display interface 15.

[0018] FIG. 2 shows a functional block diagram of a weight measurement system according to a first embodiment of the present invention. This embodiment includes a camera 26, an image acquisition unit 31, an individual information management unit 32, a planar dimension / area information acquisition unit 33, a side dimension / area information acquisition unit 34, a weight information acquisition unit 40, and a calibration unit 60. These units can be implemented by the CPU 10 appropriately reading a program from the external storage device 23 into the main memory 11, or by using a program that is initially stored in the main memory 11. Furthermore, the calibration unit 60 can be implemented by the CPU 10 appropriately reading a program from the external storage device 23 into the main memory 11, or it can be implemented by a cloud computer connected to the computer of FIG. 1 via a network. The calibration unit 60 is also included in a weight measurement system according to a second embodiment of the present invention, and will be described in detail after the weight measurement system according to the second embodiment of the present invention.

[0019] The camera 26 can be a 3D camera. The image acquisition means 31 images the target animal using the camera 26 and obtains a planar image from above and a side image from the side of the target animal. The individual information management means 32 uses the images obtained by the image acquisition means 31 to identify the target animal individually, issues and attaches identification information to the image of each individual target animal, and provides the image with this identification information attached to the planar dimension / area information acquisition means 33, the side dimension / area information acquisition means 34, and the weight information acquisition means 40.

[0020] The planar dimension / area information acquisition means 33 acquires planar dimension / area information of the target animal based on the planar image acquired by the image acquisition means. In this embodiment, the planar dimension / area information refers to planar width information and planar length information. However, it goes without saying that dimensional information, such as diagonal dimension information on a plane, or area information on a specific portion of the plane may also be used. Planar width information of the target animal will now be described. Assuming that an image of the target animal from above (here, an image facing left) is shown in U of FIG. 3, the image is divided vertically in the figure, and the width (vertical length in the figure) of the animal image is measured from the center to the left in the resulting areas. The measurement value of the largest divided area is taken as the shoulder width (FIG. 3(B)). Furthermore, measurements are also taken from the same center to the right, and the measurement value of the largest divided area is taken as the hip width (FIG. 3(B)). These can be used as planar width information. Alternatively, hip width information, etc., determined by a predetermined definition, may also be used as planar width information. The planar width information can be at least one of shoulder width information of the target animal, waist width information of the target animal, and hip width information of the target animal.

[0021] Next, the planar length information will be explained. As described above, the shoulder position and hip position are determined in the process of obtaining the planar width information, and the distance between the shoulder position and hip position is taken as the body length. This body length can be used as the planar length information. The distance between the shoulder position and hip position can also be taken as the body length.

[0022] The side dimension / area information acquisition means 34 acquires side dimension / area information of the target animal based on the side image acquired by the image acquisition means 31. In this embodiment, the side dimension / area information is body height information, but it goes without saying that dimension information in diagonal directions on the side or area information of a specific part of the side may also be used. The planar side dimension / area information of the target animal may be body height information, which is body height information of at least one of the shoulder, hip, and rump positions of the target animal. For example, as shown in FIG. 4, using a planar image U and a side image S of the target animal, first, the highest point in the longitudinal direction of the same divided area of ​​the side image S is determined. Measurements are made from the center of the image toward the left, and the height measurement of the highest divided area is defined as the shoulder height. Measurements are made from the center of the image toward the right, and the height measurement of the highest divided area is defined as the rump height. Alternatively, side dimension / area information of the hips, etc., defined according to a predetermined definition, may be used as the planar width information. The planar side dimension / area information may be withers height information, which is withers height information at least one of the shoulder position, waist position, and rump position of the target animal.

[0023] The weight information acquisition means 40 is equipped with a machine learning model 41 obtained by machine learning using the planar dimension / area information and the side dimension / area information as explanatory variables and the weight information of the target animal as a response variable. This machine learning model 41 can be created by performing machine learning using the planar dimension / area information and the side dimension / area information obtained in this embodiment as explanatory variables and actually measuring the weight information of the target animal as a response variable. The weight information acquisition means 40 provides the obtained planar dimension / area information and the side dimension / area information as explanatory variables to the machine learning model 41 to obtain weight information, which is the response variable.

[0024] The weight measurement system configured as described above performs processing according to a program corresponding to the flowchart shown in Fig. 5. The operation will be described below based on the flowchart. The target animal is imaged by the camera 26 (S11). The process of imaging the target animal by the camera 26 continues in the processes in the following steps.

[0025] When the target animal is imaged by the camera 26, the CPU 10 receives the image data (S12) and calculates the planar dimension and width information and the lateral dimension and width information (S13). The CPU 10 provides the calculated planar dimension and width information and the lateral dimension and width information to a machine learning model to obtain the weight of the target animal (S14). In this way, it becomes possible to appropriately measure the weight of the target animal.

[0026] Next, a weight measurement system according to a second embodiment will be described. Fig. 6 shows a functional block diagram of a weight measurement system according to a second embodiment of the present invention. In this embodiment, the camera 26 and image acquisition means 31 are the same as those in the first embodiment. This embodiment also includes individual information management means 32, individual target animal image acquisition means 35, appropriate image selection means 37, and correction processing means 38. Furthermore, in this embodiment, the target animal will be described as a pig, but the target animal of the present invention is not limited to this.

[0027] The individual information management means 32 performs individual identification (image recognition) of the target animals using the images obtained by the image acquisition means 31, assigns identification information to the images of each individual target animal, and provides the images with this identification information to the planar dimension / area information acquisition means 33, the lateral dimension / area information acquisition means 34, and the weight information acquisition means 40. In other words, when multiple target animals are kept together, the individual information management means 32 assigns identification information to each target animal and manages the images. More specifically, it uses RGB images for tracking and generates a bounding box for each pig. When there are multiple pigs, a separate bounding box and its coordinates are calculated for each pig, and the results are passed to the individual target animal image acquisition means 35, which is a target selection unit.

[0028] This embodiment is suitable for cases where multiple target animals are kept together, and the camera 26 captures an image of the location where the multiple target animals are present. As a result, the multiple target animals appear overlapping in the image. The individual target animal image acquisition means 35 separates the images of the individual target animals from the image captured by the camera 26 to acquire the individual target animal images. In this embodiment, a segmentation method can be employed. By using this segmentation method, even overlapping pigs can be clearly separated, and a mask image of each pig is generated and used for processing by the weight information acquisition means 40.

[0029] The appropriate image selection means 37 performs processing to select only images that can be processed appropriately when providing images with the identification information to the planar dimension / area information acquisition means 33, the side dimension / area information acquisition means 34, and the weight information acquisition means 40. If the pig is in a bent position, the values ​​of each part cannot be measured correctly. Therefore, we considered what to do with images in which the body is bent. Since this embodiment uses a fixed camera, it is possible to photograph the same pig multiple times, so we decided to discard data when the body is bent and use only images taken in a good position.

[0030] Therefore, we decided to eliminate images in which the back was bent. Next, we can determine whether the body is bent as follows. When obtaining planar dimension and width information, multiple divided areas are created as shown in Figure 3, and the line connecting the highest points of each divided area in Figure 7 corresponds to the spine. Detecting the degree of curvature of this spine can determine the pig's posture. An ellipse is fitted in advance using image processing, and the center line of this ellipse is the ideal position of the spine. Images that deviate significantly from this line are judged to have high curvature and are eliminated (Figure 7(B)), while images that deviate less from this line are retained (7(A)). For example, images that are more than a certain distance from the line connecting the highest points of each divided area can be eliminated. Furthermore, because pigs judged to be normal are also slightly bent, data on the deviation from the ideal position indicating the curvature of the spine (e.g., variance) can also be added as an explanatory variable.

[0031] The correction processing means 38 performs correction processing to orient the image selected by the appropriate image selection means 37 in a predetermined direction. In this embodiment, measurements of each part are performed by the planar dimension / area information acquisition means 33 and the lateral dimension / area information acquisition means 34 using 3D images of one pig extracted by the individual target animal image acquisition means 35 using a segmentation technique. The parts to be measured can include shoulder width information, waist width information, hip width information, length from shoulder to hip, and total body area. To measure each part, it is easy to calculate the measurements by cutting the pig image into slices at regular intervals and processing them sequentially from the head, as described above. However, in this embodiment, images are taken of pigs coming to eat, so images are acquired in various orientations. Since the posture is not constant, some ingenuity is required. The processing performed by the correction processing means 38, including solutions to these issues, is described below.

[0032] <Correction processing> The 3D image is binarized and an ellipse is approximated along the boundary of the body. If the image of the pig is cut out cleanly, an ellipse that fits the body can be drawn, with the center of the ellipse corresponding to the center of the body and the angle of the ellipse matching the angle of the pig. Therefore, the angle of the ellipse in the state shown in Figure 8(A) is calculated, and the image is rotated using an affine transformation with the center of the ellipse as the rotation axis so that the angle of the ellipse that makes the image horizontal is horizontal (Figure 8(B)).

[0033] The weight measurement system configured as described above performs processing according to a program corresponding to the flowchart shown in Fig. 9. The operation will be explained below based on the flowchart. CPU 10 captures an image of the target animal with camera 26 (S11). Once the target animal is captured by camera 26, CPU 10 receives the captured image data (S12) and separates overlapping images of the target animals into a single image (S21).

[0034] Next, CPU 10 performs individual identification (image recognition) and assigns identification information to the image of each individual target animal (S22). Next, distorted images are discarded and only images captured in good condition are used (S23). CPU 10 also performs ellipse approximation along the boundary surface of the target animal's body, calculates the angle of the ellipse, and performs correction by rotating the image using affine transformation with the center of the ellipse as the axis of rotation so that the angle of the ellipse is horizontal (S24). Furthermore, planar dimension and width information, and lateral dimension and width information are calculated (S13). CPU 10 provides the calculated planar dimension and width information and lateral dimension and width information to a machine learning model to obtain the weight of the target animal (S14). This makes it possible to accurately measure the weight of the target animal.

[0035] Furthermore, the weight information acquisition means 40 may calculate the average value of weight information obtained within a predetermined time period for a target animal identified as an individual, and use the calculated average value as the weight of the target animal. Even if the target animal remains within the field of view, there may be cases where weight estimation is not possible due to poor posture or other conditions, resulting in a reduced number of weight measurements. If the acquired weight measurement value is below a set threshold, the acquired data may be discarded. Only when weight measurements are obtained the threshold number of times or more may the data be averaged, and the weight measurement value for one target animal may be completed.

[0036] Furthermore, the weight measurement system may be provided with a program realizing a type determination means (not shown) that determines type information within the target animal based on the planar dimension / area information and the side dimension / area information. In this case, the individual information management means 35 performs an operation of providing an image with identification information attached together with the type information determined by the type determination means to the planar dimension / area information acquisition means 33, the side dimension / area information acquisition means 34, and the weight information acquisition means 40, and the weight information acquisition means 40 may adjust the weight information using parameters according to the type information within the target animal when acquiring weight information.

[0037] The above-described embodiments can reduce the labor required for measuring pig weight, enable daily management and data collection of pig weight, and can also lead to maximization of profits for pig farms by optimizing shipping times.

[0038] The configurations of the weight measurement system of the first embodiment and the weight measurement system of the second embodiment described above raise the following technical issues. That is, will the accuracy of these embodiments pose any problems in actual operation? Furthermore, can the same accuracy be ensured when the system of this embodiment is deployed to various farms? Furthermore, can it be operated with the same accuracy even if the type of pig changes? In consideration of these issues, the accuracy of the machine learning model 41 at a certain time was estimated, as shown in Figure 13. That is, the root mean square error (RMSE) was 2.6%. From this perspective, the weight measurement system of the first embodiment and the weight measurement system of the second embodiment of the present invention are equipped with a calibration unit 60. The calibration unit 60 will be described below.

[0039] The calibration unit 60 includes a history information storage means 61, an updated learning model generation means 62, and an update timing control means 63. The history information storage means 61 stores pairs of explanatory variables and objective variables as history information when the weight information acquisition means 40 acquires weight information using the machine learning model 41. The pairs of explanatory variables and objective variables are stored in association with identification information issued for each individual by the individual information management means 32.

[0040] The updated learning model generation means 62 acquires an updated machine learning model by performing the same machine learning as that used to create the machine learning model 41, based on input of actual weight information of the target animal corresponding to the explanatory variables in the pair of historical information stored in the historical information storage means 61, using the actual weight information as an explanatory variable and the actual weight information as the objective variable. More specifically, an updated machine learning model may be acquired when a predetermined number of pieces of actual weight information serving as the objective variable are obtained. That is, for example, the weight of the target animal can be measured using a weighing scale at the time of shipping, and the weight can be input to the calibration unit 60 in association with the identification information to create pairs of objective variables and explanatory variables to be used in generating the updated machine learning model. Alternatively, identification information from a butcher shop and weight information obtained by measuring the corresponding meat, known as a carcass, obtained by processing one animal can be used.

[0041] Furthermore, if a sufficient number of pairs of dependent variables and explanatory variables to be used in generating an updated machine learning model cannot be obtained, a combination of dependent variable and explanatory variable pairs created from actual weight information and pairs of dependent variables and explanatory variables used in generating machine learning model 41 may be used. Furthermore, if weight information acquisition means 40 adjusts weight information using parameters according to the species information of the target animal, calibration may be performed to change these parameters according to actual weight information. Furthermore, calibration of machine learning model 41 of the farm system to be calibrated may be performed using pairs of dependent variables and explanatory variables created from actual weight information obtained from another farm with a similar breeding method and pig species ratio.

[0042] An update timing control means 63 is provided to determine the error between the weight information and actual weight information obtained using the machine learning model. The update timing control means 63 creates error information between the weight information and actual weight information obtained using the machine learning model 41, and controls the timing of obtaining an updated machine learning model based on this error information. Therefore, the update timing control means 63 can be used to create and output error information between the weight information and actual weight information obtained using the machine learning model 41, allowing the error to be monitored.

[0043] The update timing control means 63 can notify that it is time to update when the root mean square error (RMSE) shown in Fig. 13 exceeds, for example, 3%. Alternatively, the updated learning model generation means 62 may be caused to operate using pairs of objective variables and explanatory variables created from collected actual body weight information, and the generated updated learning model may be used as the new machine learning model 41.

[0044] The calibration unit 60 is expected to be able to appropriately address issues such as whether the accuracy of this embodiment will pose any problems in actual operation, whether similar accuracy can be ensured when the system of this embodiment is deployed to various farms, and whether the system can be operated with the same accuracy even if the type of pig changes. [Explanation of symbols]

[0045] 10 CPU 11 Main Memory 12 Bus 13 External memory interface 14 Input Interface 15 Display Interface 16 Data Input Interface 22 Pointing Device 23 External storage device 24 Input Devices 25 Display device 26 Camera 31 Image acquisition means 32 Individual information management means 33 Method of obtaining floor plan dimensions and area information 34 Means for obtaining side dimensions and width information 35 Individual target animal image acquisition means 37 Appropriate image selection method 38 Correction processing means 40 Weight information acquisition means 41 Machine Learning Models 60 Calibration section 61 History information storage means 62 Update learning model generation means 63 Update timing control means

Claims

1. an image acquisition means for capturing an image of a target animal using a camera to obtain a planar image from above and a side image from the side of the target animal; a planar size / area information acquisition means for acquiring planar size / area information of the target animal based on the planar image; side dimension / area information acquisition means for acquiring side dimension / area information of the target animal based on the side image; a weight information acquisition means for providing the planar dimension / area information and the side dimension / area information as explanatory variables to the machine learning model, the weight information being the objective variable, and the planar dimension / area information and the side dimension / area information as explanatory variables to the machine learning model; and a calibration unit that calibrates the machine learning model; A weight measurement system comprising:

2. The calibration unit a history information storage means for storing pairs of explanatory variables and objective variables as history information when the weight information acquisition means acquires weight information using the machine learning model; an updated learning model generation means for performing the same machine learning as when the machine learning model was created, based on input of actual body weight information of the target animal corresponding to an explanatory variable in a pair in the history information stored in the history information storage means, using the actual body weight information as an explanatory variable and the actual body weight information as a target variable, to obtain an updated machine learning model; 2. The weight measurement system according to claim 1, further comprising:

3. 3. The weight measurement system according to claim 2, wherein the actual weight information is obtained by measuring the weight of the target animal using a weighing scale.

4. 3. The weight measurement system according to claim 2, wherein the actual weight information is weight information obtained by measuring meat obtained by processing the target animal.

5. The calibration unit includes: The weight measurement system according to claim 2, further comprising an update timing control means for creating error information between the weight information acquired using the machine learning model and the actual weight information, and controlling the timing for acquiring an updated machine learning model based on this error information.

6. The weight measurement system according to claim 1, further comprising an individual information management means for performing individual identification of the target animal using the image obtained by the image acquisition means, attaching identification information to the image of each individual target animal, and providing the image with this identification information attached to the planar dimension / area information acquisition means, the lateral dimension / area information acquisition means, and the weight information acquisition means.

7. the camera captures an image of a location where a plurality of the target animals are present; 2. The weight measurement system according to claim 1, further comprising an individual target animal image acquisition means for acquiring individual target animal images by separating images of the individual target animals from the image obtained by the camera.

8. The weight measurement system according to claim 6, further comprising an appropriate image selection means for selecting only images that can be processed appropriately when the image with the identification information attached is provided to the planar dimension / area information acquisition means, the side dimension / area information acquisition means, and the weight information acquisition means.

9. 9. The weight measurement system according to claim 8, further comprising a correction processing unit that performs a correction process to orient the image selected by the appropriate image selection unit in a predetermined direction.

10. 2. The weight measurement system according to claim 1, wherein the planar width information of the target animal is at least one of shoulder width information of the target animal, waist width information of the target animal, and hip width information of the target animal.

11. The weight measurement system according to claim 1 , wherein the planar length information of the target animal is a body length, which is the length from the shoulder to the hip of the target animal.

12. 2. The weight measurement system according to claim 1, wherein the planar side dimension / area information of the target animal is withers height information of at least one of the shoulder position, waist position, and hip position of the target animal.

13. The weight measurement system according to claim 6, characterized in that the weight information acquisition means calculates the average value of weight information obtained within a predetermined time period for a target animal identified as an individual, and sets the calculated average value as the weight of the target animal.

14. a species determination means for determining species information within the target animal based on the planar dimension / area information and the side dimension / area information; the individual information management means performs an operation of providing the type information determined by the type determination means and an image with the type information attached to the planar dimension / area information acquisition means, the side dimension / area information acquisition means, and the weight information acquisition means; 7. The weight measurement system according to claim 6, wherein the weight information acquisition means adjusts the weight information by using a parameter corresponding to the type information when acquiring the weight information.

15. an image acquisition step of capturing an image of a target animal using a camera to obtain a planar image from above and a side image from the side of the target animal; a planar size / area information acquisition step of acquiring planar size / area information of the target animal based on the planar image; a side dimension / area information acquisition step of acquiring side dimension / area information of the target animal based on the side image; a weight information acquisition step of using a machine learning model obtained by machine learning with the planar dimension / area information and the side dimension / area information as explanatory variables and the weight information of the target animal as a target variable, and providing the planar dimension / area information and the side dimension / area information obtained as explanatory variables to the machine learning model to obtain weight information as a target variable; a history information accumulation step of accumulating pairs of explanatory variables and objective variables as history information when weight information is acquired using the machine learning model in the weight information acquisition step; an updated learning model generation step of performing the same machine learning as when the machine learning model was created, based on input of actual body weight information of the target animal corresponding to an explanatory variable in a pair in the history information accumulated in the history information accumulation step, using the actual body weight information as an explanatory variable and the actual body weight information as a target variable, to obtain an updated machine learning model; A weight measurement method comprising:

16. 16. The weight measurement method according to claim 15, wherein the actual weight information is obtained by measuring the weight of the subject animal using a weighing scale.

17. The weight measurement method according to claim 15, wherein the actual weight information is weight information obtained by measuring meat obtained by processing the target animal.

18. The weight measurement method described in claim 15, characterized in that it includes an update timing control step that creates error information between the weight information obtained using the machine learning model and the actual weight information, and controls the timing of obtaining an updated machine learning model based on this error information.

19. The weight measurement method according to claim 15, further comprising an individual information management step of performing individual identification of the target animal using the image obtained by the image acquisition step, attaching identification information to the image of each individual target animal, and providing the image with this identification information attached to the planar dimension / area information acquisition step, the lateral dimension / area information acquisition step, and the weight information acquisition step.

20. the camera captures an image of a location where a plurality of the target animals are present; The weight measurement method according to claim 15, further comprising an individual target animal image acquisition step of acquiring individual target animal images by separating images of the individual target animals from the image obtained by imaging with the camera.

21. The weight measurement method according to claim 19, further comprising an appropriate image selection step for selecting only images that can be processed appropriately when the image with the identification information attached is provided to the planar dimension / area information acquisition step, the side dimension / area information acquisition step, and the weight information acquisition step.

22. 22. The weight measurement method according to claim 21, further comprising a correction processing step of performing a correction process to orient the image selected in the appropriate image selection step so that the image is oriented in a predetermined direction.

23. 16. The weight measurement method according to claim 15, wherein the planar width information of the target animal is at least one of shoulder width information of the target animal, waist width information of the target animal, and hip width information of the target animal.

24. The weight measurement method according to claim 15, wherein the planar length information of the target animal is the length from the shoulder to the hip of the target animal.

25. 16. The weight measuring method according to claim 15, wherein the planar side dimension / area information of the target animal is withers height information of at least one of the shoulder position, waist position, and hip position of the target animal.

26. The weight measurement method described in claim 19, characterized in that the weight information acquisition step calculates the average value of weight information obtained within a predetermined time for a target animal identified as a single individual, and sets the calculated average value as the weight of the target animal.

27. a type determination step of determining type information of the target animal based on the planar dimension / area information and the side dimension / area information, the individual information management step performs an operation of providing the type information determined in the type determination step and an image with the type information attached to the planar dimension / area information acquisition step, the side dimension / area information acquisition step, and the weight information acquisition step; 20. The weight measurement method according to claim 19, wherein the weight information acquisition step performs adjustment using parameters corresponding to the type information when obtaining the weight information.

Images