Body Gain Assessment System, Body Gain Assessment Method, and Program

Abstract

To provide a technique capable of determining weight gains of pigs.SOLUTION: An estimation system according to one aspect of the present disclosure is a weight gain determination system including an imaging device for imaging domestic animals, and a weight gain determination device for determining weight gains of the domestic animals. The weight gain determination device includes a chest width estimation part for estimating chest widths of the domestic animals, on the basis of imaging data acquired from the imaging device, and a weight gain determination part for determining the weight gain of the domestic animal, on the basis of the chest width estimated by the chest width estimation part, and a standard weight gain curve indicating a relation between the number of elapsed days from an initial mating day of the domestic animal and a standard chest width.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a weight gain determination system, a weight gain determination method, and a program. [Background technology]

[0002] In the field of pig farming, an index called the body condition score (BCS), which is expressed on a five-point scale, has been known for some time. The body condition score is an index that indicates the degree of obesity of pigs, and is often used to understand the weight gain of breeding sows and to manage their nutrition. Because the nutritional state of breeding sows affects their reproductive performance, it is important to accurately understand their body condition score. For this reason, methods for determining the body condition score have been known for some time (for example, Non-Patent Document 1). [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Motoo Iguchi, "Using Body Condition Score as a Guide to Feed Amount for Breeding Sows - Improving Reproductive Performance is All About Controlling Body Condition!", Production Technology Seminar Summary of the Invention [Problem to be solved by the invention]

[0004] However, the conventional five-level body condition score is determined subjectively based on the appearance and palpation of the sow, and as a result, the accuracy of the score determination is low depending on the experience of the scorer, and as a result, the weight gain of the sow may not be accurately understood.

[0005] The present disclosure has been made in consideration of the above points and provides a technology that can determine the weight gain of pigs. [Means for solving the problem]

[0006] An estimation system according to one aspect of the present disclosure is a weight gain determination system that includes a photographing device that photographs livestock and a weight gain determination device that determines the weight gain of the livestock, wherein the weight gain determination device has a chest width estimation unit that estimates the chest width of the livestock based on photographing data obtained from the photographing device, and a weight gain determination unit that determines the weight gain of the livestock based on the chest width estimated by the chest width estimation unit and a standard weight gain curve that represents the relationship between the number of days elapsed since the first mating date of the livestock and a standard chest width that represents a standard chest width. [Effects of the Invention]

[0007] A technique is provided that can determine the weight gain of pigs. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a diagram illustrating an example of the overall configuration of a body weight gain determination system according to an embodiment of the present invention. [Figure 2] FIG. 2 is a diagram illustrating an example of the functional configuration of the imaging device according to the present embodiment. [Figure 3] FIG. 2 is a diagram illustrating an example of a functional configuration of a body growth determination device according to the present embodiment. [Figure 4] 10 is a flowchart illustrating an example of a photographing process according to the present embodiment. [Figure 5] FIG. 10 is a diagram illustrating an example of a shooting screen. [Figure 6] 10 is a flowchart illustrating an example of a body weight gain determination process according to the present embodiment. [Figure 7] FIG. 1 is a diagram showing an example of a standard weight gain curve. [Figure 8] 10 is a flowchart showing an example of a standard body weight gain curve generation process according to the present embodiment. [Figure 9] FIG. 1 shows an example of a baseline weight gain curve and an adjusted weight gain curve. DETAILED DESCRIPTION OF THE INVENTION

[0009] An embodiment of the present invention will be described below. In the following embodiment, a weight gain determination system 1 will be described that is targeted at breeding sows and can estimate the chest width of the breeding sow and then determine the weight gain from the chest width.

[0010] The reason for estimating chest width when determining weight gain is that if the chest width of a breeding sow can be accurately estimated, then it will also be possible to accurately determine her weight gain. This is because weight gain is an index that represents the degree of obesity, nutritional status, etc., and it is generally known that there is a strong correlation between the degree of obesity, nutritional status, etc., and body weight. Furthermore, chest width naturally has a strong correlation with (or can be considered identical to) chest circumference, and it is known that there is a strong correlation between the weight and chest circumference of a pig (see, for example, Reference 1).

[0011] <Overall configuration example of body weight gain assessment system 1> An example of the overall configuration of a body weight gain determination system 1 according to this embodiment is shown in Fig. 1. As shown in Fig. 1, the body weight gain determination system 1 according to this embodiment includes a photographing device 10 that photographs a breeding sow P, whose body weight gain is to be determined, and a body weight gain determination device 20 that estimates the chest width of the breeding sow P and then determines the body weight gain from the chest width. Furthermore, the photographing device 10 and the body weight gain determination device 20 are connected to each other so that they can communicate with each other, for example, wirelessly or by wire, or both.

[0012] The photographing device 10 is, for example, a smartphone, tablet terminal, digital camera, or the like used by a user M, such as a manager or employee of a pig farm. The user M uses the photographing device 10 to photograph a breeding sow P in a gestation stall C (hereinafter simply referred to as "stall C") from above. The photographing device 10 may be a camera fixedly installed above the stall C.

[0013] Here, the photographing device 10 is equipped with at least an infrared camera (or infrared sensor) and a depth sensor (or depth camera), and after user M starts photographing, an infrared image and point cloud data within the photographing range are generated if predetermined photographing conditions are met. Point cloud data is data expressed as a cloud of points (x, y, z) where the horizontal direction within the photographing range is the x-axis, the vertical direction is the y-axis, and the depth direction is the z-axis. Point cloud data may also be called "three-dimensional point cloud data," etc. User M photographs breeding sow P so that at least the area of ​​the back of breeding sow P, from the shoulders to the chest, is included within the photographing range, and so that the spine of breeding sow P is as parallel as possible to the vertical direction (y-axis direction) of the photographing range.

[0014] The weight gain determination device 20 is a computer or computer system that estimates the chest width of the breeding sow P from the infrared image and point cloud data generated by the photographing device 10, and then determines the weight gain from the estimated chest width using a graph called a standard weight gain curve, which shows the relationship between the number of days elapsed since the first mating date and the standard chest width.

[0015] The overall configuration of the body weight gain determination system 1 shown in Fig. 1 is an example and is not limited to this. For example, the body weight gain determination system 1 shown in Fig. 1 may include devices, apparatuses, terminals, etc. other than the photographing device 10 and the body weight gain determination device 20. Furthermore, for example, the photographing device 10 and the body weight gain determination device 20 may be configured as an integrated unit.

[0016] <Example of functional configuration of the imaging device 10> An example of the functional configuration of the image capturing device 10 according to this embodiment is shown in Fig. 2. As shown in Fig. 2, the image capturing device 10 according to this embodiment includes a transmission / reception unit 101, a UI control unit 102, an image capturing condition determination unit 103, an image capturing unit 104, and an image capturing data creation unit 105. Each of these units is realized by, for example, processing executed by a calculation device such as a CPU (Central Processing Unit) by one or more programs installed in the image capturing device 10.

[0017] The transmitting / receiving unit 101 transmits the imaging data created by the imaging data creating unit 105 to the body weight gain determination device 20, and receives body weight gain determination result data representing the body weight gain determination result from the body weight gain determination device 20.

[0018] The UI control unit 102 displays various screens (e.g., the shooting screen described below) on the display of the shooting device 10 and accepts user operations on these various screens (e.g., a setting operation for setting the number of days that have passed since the first mating date of the breeding sow P to be photographed, an operation to start shooting to display the shooting screen, etc.).

[0019] When the UI control unit 102 receives an operation to start shooting by the user M, the shooting condition determination unit 103 determines at predetermined time intervals (for example, at each depth measurement cycle) whether or not predetermined shooting conditions are met. The shooting conditions are conditions for determining whether the relative positional relationship between the shooting device 10 and the breeding sow P within the shooting range, the shooting environment, etc. are appropriate. Details of the shooting conditions will be described later.

[0020] If the shooting condition determination unit 103 determines that the shooting conditions are met, the shooting unit 104 captures the shooting range using an infrared camera (or infrared sensor) and a depth sensor (or depth camera), and generates an infrared image and point cloud data within the shooting range.

[0021] The photographing data creation unit 105 creates photographing data that includes the infrared image and point cloud data generated by the photographing unit 104 and the number of days that have passed since the first mating date of the breeding sow P that was the subject of the photographing (i.e., the number of days that have passed since the first mating date of the breeding sow P that was the subject of the photographing).

[0022] <Example of functional configuration of body weight gain determination device 20> An example of the functional configuration of the body weight gain determination device 20 according to this embodiment is shown in Fig. 3. As shown in Fig. 3, the body weight gain determination device 20 according to this embodiment includes a transmitter / receiver 201, a chest width estimation unit 202, a body weight gain determination unit 203, a body weight gain determination result data creation unit 204, and a standard body weight gain curve generation unit 205. Each of these units is realized, for example, by a process in which one or more programs installed in the body weight gain determination device 20 are executed by a computing device such as a CPU. Note that some of these units may be realized, for example, by functions provided by a cloud service or the like.

[0023] The body weight gain determination device 20 according to this embodiment also includes a storage unit 206. The storage unit 206 is realized by a storage device such as a hard disk drive (HDD) or a solid state drive (SSD). The storage unit 206 may also be realized by a storage device (e.g., a database server) connected to the body weight gain determination device 20 via a communication network.

[0024] The transmitting / receiving unit 201 receives imaging data from the imaging device 10 and transmits body weight gain determination result data created by the body weight gain determination result data creating unit 204 to the imaging device 10.

[0025] The chest width estimation unit 202 estimates the chest width of the breeding sow P that was the subject of the image capture, using the infrared image and point cloud data included in the image capture data received by the transmission / reception unit 201. At this time, the chest width estimation unit 202 uses the point cloud data to determine the number of pixels in the infrared image per predetermined unit width, and estimates the chest width from this number of pixels. This makes it possible to accurately estimate the chest width of the breeding sow P that was the subject of the image capture, even if there is a gap in part of the point cloud data. Note that depth sensors (or depth cameras) are generally sensitive to brightness, and gaps in part of the point cloud data may occur in an environment with a large difference in light and dark.

[0026] The weight gain determination unit 203 determines the weight gain of the breeding sow P that was the subject of the photograph using the standard weight gain curve stored in the memory unit 206, the chest width estimated by the chest width estimation unit 202, and the number of days elapsed included in the photographed data received by the transmission / reception unit 201.

[0027] The body weight gain determination result data creating unit 204 creates body weight gain determination result data including the body weight gain determination result determined by the body weight gain determining unit 203 .

[0028] The standard gain curve generating unit 205 generates a standard gain curve using the performance data set stored in the storage unit 206 and the reference gain curve. Here, the performance data set is a collection of performance data including the number of days elapsed since the breeding sow's birth and its chest width. This performance data is generated, for example, by a user M or the like actually measuring the breeding sow's chest width and associating it with the number of days elapsed since the breeding sow's first mating date. The reference gain curve is a graph showing the relationship between the number of days elapsed since the first mating date and a typical chest width that serves as a standard for that number of days elapsed. In other words, the reference gain curve is a graph showing the relationship between the number of days elapsed since the first mating date and the chest width of a typical pig at that number of days elapsed.

[0029] The storage unit 206 stores various data. For example, when generating a standard weight gain curve, a performance data set and a reference weight gain curve are stored in the storage unit 206. Furthermore, for example, when determining the weight gain of a breeding sow P, the storage unit 206 stores the standard weight gain curve generated by the standard weight gain curve generation unit 205.

[0030] <Shooting process> The imaging process according to this embodiment will be described below with reference to Fig. 4. In the following, it is assumed that the storage unit 206 of the body weight gain determination device 20 stores a standard body weight gain curve.

[0031] When the user M performs an operation to set the number of days that have passed (that is, the number of days that have passed since the first mating date of the breeding sow P to be photographed), the UI control unit 102 accepts this setting operation (step S101).

[0032] Next, when a shooting start operation is performed by the user M, the UI control unit 102 accepts this shooting start operation (step S102).

[0033] Next, when the UI control unit 102 receives an operation to start shooting, it displays a shooting screen on the display (step S103). An example of the shooting screen displayed by the UI control unit 102 is shown in Fig. 5. The shooting screen 1000 shown in Fig. 5 includes a shooting guide 1100 for adjusting the shooting position of the breeding sow P. The shooting guide 1100 has a shape that represents the outline of the pig when viewed from above, with the downward direction of the shooting screen 1000 being the head direction and the upward direction being the tail direction.

[0034] The shooting screen 1000 also includes a center guide 1110 that indicates the center of the shooting range 1200. In this embodiment, the origin of the point cloud data (i.e., the position where (x, y, z) = (0, 0, 0)) is the camera position of the shooting device 10. Therefore, the position indicated by the center guide 1110 is the origin of an arbitrary xy plane of the point cloud data (i.e., (x, y) = (0, 0)). The shooting direction is the positive direction of the z axis, the downward direction of the shooting screen 1000 is the positive direction of the x axis, and the rightward direction of the shooting screen 1000 is the positive direction of the y axis. The x axis is also the center line that bisects (equally divides) the shape represented by the shooting guide 1100 into left and right halves.

[0035] The user M moves the photographing device 10 above the breeding sow P and adjusts it so that the outline of the breeding sow P matches the photographing guide 1100. This allows the breeding sow P to be photographed when the photographing conditions described below are met.

[0036] The photography screen 1000 may include an inversion button 1300 for inverting the photography guide 1100 upside down. This inversion button 1300 makes it possible to change the upside down of the shape represented by the photography guide 1100 (the shape representing the outline of the pig when viewed from above) in accordance with the actual orientation of the breeding sow P.

[0037] The photographing screen 1000 may also include a photographing button 1400 for photographing the area within the photographing range 1200. This photographing button 1400 can be used to photograph the breeding sow P manually, as well as when the photographing conditions described below are met. Hereinafter, the explanation of FIG. 4 will be continued assuming that the photographing screen 1000 shown in FIG. 5 is displayed on the display of the photographing device 10.

[0038] Following step S103, the photographing condition determination unit 103 determines whether or not predetermined photographing conditions are met (step S104). The photographing conditions are conditions for determining whether the relative positional relationship between the photographing device 10 and the breeding sow P within the photographing range 1200, the photographing environment, etc. are appropriate, and for example, at least one of the following conditions 1 to 5 can be used.

[0039] Condition 1: The depth value at the center of the photographing range 1200 (i.e., the z coordinate value at (x, y) = (0, 0)) is less than a predetermined threshold. This is a condition for easily checking whether the breeding sow P is within the photographing range.

[0040] Condition 2: The posture of the breeding sow P estimated from the increase / decrease pattern of the depth values ​​of the point cloud data on the x-axis of the shooting range 1200 and axes parallel to it (i.e., depth values ​​on any axis parallel to the x-axis) is not a predetermined posture. This is because the increase / decrease pattern of the depth values ​​on an axis parallel to the x-axis makes it possible to estimate postures such as head-up, head-down, or knee-bent postures of the breeding sow P, and these postures may reduce the accuracy of the chest width estimation.

[0041] If the breeding sow P's head is positioned downward within the shooting range 1200, the depth value decreases as the distance from the origin increases in the positive direction on an axis parallel to the x-axis in a head-up posture, and this depth value pattern can be used to estimate the head-up posture. Similarly, the depth value increases as the distance from the origin increases in the positive direction on an axis parallel to the x-axis in a head-down posture, and this depth value pattern can be used to estimate the head-down posture. Similarly, the depth value increases as the distance from the origin increases in the negative direction on an axis parallel to the x-axis in a knee-bent posture, and this depth value pattern can be used to estimate the knee-bent posture.

[0042] Condition 3: The position of stall C (more precisely, the iron bars and the like that make up stall C) estimated from the pattern of increase and decrease in depth values ​​on the x-axis of the photographing range 1200 and on an axis parallel to it is not the position of the chest of breeding sow P. This is because if the position of the chest of breeding sow P overlaps with the position of stall C, it is not possible to estimate the chest width.

[0043] Since the depth value on the axis parallel to the x-axis on stall C remains almost constant, the position of stall C can be estimated from the pattern of depth values. Furthermore, for example, if the absolute value of the difference between the x-coordinate value of the position of stall C and a predetermined value as the x-coordinate value of the position of the pig's chest is less than a predetermined threshold, it can be considered that the position of the breeding sow P's chest and the position of stall C overlap.

[0044] Condition 4: The body length of the breeding sow P estimated from the pattern of increase and decrease in depth values ​​on the x-axis and the axis parallel to it of the photographing range 1200 is equal to or greater than a predetermined threshold. This is a condition for preventing the photographing device 10 from being too far away from the breeding sow P.

[0045] When moving in the positive direction from the origin on an axis parallel to the x-axis, the depth value increases sharply at the boundary of the breeding sow P. Similarly, when moving in the negative direction from the origin on an axis parallel to the x-axis, the depth value also increases sharply at the boundary of the breeding sow P. This allows the body length of the breeding sow P to be estimated.

[0046] Condition 5: In the histogram of infrared camera (or infrared sensor) values ​​within the shooting range 1200, the total number of frequencies below a certain specified class is less than a predetermined threshold. This is a condition for checking the brightness of the shooting environment, and if this condition is met, it indicates that the brightness of the shooting environment is sufficient.

[0047] When Intel (registered trademark) RealSense is used as the depth sensor (or depth camera) of the imaging device 10, under the above condition 2, the (x, y, z) coordinate values ​​of the area of ​​the breeding sow P may be extracted using a function called Threshold Filter provided in RealSense, and then imaged using Colorizer, and the posture may be estimated from the shape of the area represented by this image. Similarly, under the above condition 3, the (x, y, z) coordinate values ​​of the area of ​​stall C (more precisely, the iron bars and the like that make up stall C) may be extracted, and then imaged using Colorizer, and the position of stall C may be estimated from the position of the area represented by this image.

[0048] By using at least one of the above conditions 1 to 5 as the photographing conditions, it becomes possible to photograph automatically, for example, when the relative positional relationship with the breeding sow P is appropriate or when the photographing environment is appropriate.

[0049] If it is determined in step S104 above that the photographing conditions are not met, the photographing device 10 returns to step S104 (NO in step S105). On the other hand, if it is determined in step S104 above that the photographing conditions are met, the photographing device 10 proceeds to step S106 (YES in step S105).

[0050] The photographing unit 104 photographs the inside of the photographing range 1200 using an infrared camera (or an infrared sensor) and a depth sensor (or a depth camera) (step S106). As a result, an infrared image of the inside of the photographing range 1200 and point cloud data of the inside of the photographing range are generated by the photographing unit 104.

[0051] Next, the shooting data creation unit 105 creates shooting data including the infrared image and point cloud data generated in the above step S106 and the number of days that have passed set by the setting operation accepted in the above step S101 (step S107).

[0052] Next, the transmitting / receiving unit 101 transmits the photographed data created in step S107 to the body weight gain determination device 20 (step S108). As a result, a body weight gain determination process, which will be described later, is executed by the body weight gain determination device 20, and as a result, body weight gain determination result data is transmitted from the body weight gain determination device 20 to the photographing device 10.

[0053] Next, the transmitting / receiving unit 101 receives the body weight gain determination result data transmitted from the body weight gain determination device 20 (step S109).

[0054] Then, the UI control unit 102 displays the weight gain determination result included in the weight gain determination result data received in step S109 on the display (step S110). For example, if the weight gain determination result is information representing one of the five categories of "too thin," "slightly thin," "normal," "slightly overweight," or "too overweight," the UI control unit 102 displays letters, figures, symbols, etc. representing one of "too thin," "slightly thin," "normal," "slightly overweight," or "too overweight" on the display. This allows the user M to know whether the weight gain of the breeding sow P that was the subject of the photograph is "too thin," "slightly thin," "normal," "slightly overweight," or "too overweight," compared to a standard weight gain.

[0055] <Body Gain Determination Processing> The body weight gain determination process according to this embodiment will be described below with reference to Fig. 6. In the following, it is assumed that the storage unit 206 of the body weight gain determination device 20 stores a standard body weight gain curve.

[0056] The transmitting / receiving unit 201 receives the image capturing data transmitted from the image capturing device 10 (step S201). The image capturing data includes an infrared image, point cloud data, and the number of days that have passed.

[0057] Next, the chest width estimation unit 202 estimates the chest width of the breeding sow P that was the subject of the photography using the infrared image and point cloud data included in the photography data received in step S201 above (step S202). At this time, the chest width estimation unit 202 estimates the chest width, for example, by the following steps 1 to 5. Note that in the point cloud data, the depth value (z coordinate value) represents the distance from the photography device 10, so it is possible to easily calculate the distance between any two points whose depth values ​​are known.

[0058] Step 1: The chest width estimation unit 202 aligns each point group (x, y, z) represented by the point cloud data at intervals of a predetermined unit width (for example, 1 cm) in the y-axis direction. In other words, the chest width estimation unit 202 creates point cloud data aligned at intervals of a predetermined unit width in the y-axis direction by deleting a part of each point group (x, y, z) represented by the point cloud data or by interpolating the point group. In the following, the explanation will be continued assuming that the unit width is 1 cm as an example.

[0059] Step 2: The chest width estimation unit 202 visualizes the point cloud data (point cloud data aligned at 1 cm intervals in the y-axis direction) created in step 1 above into an image of the same size as the infrared image. Hereinafter, the visualized point cloud data will be referred to as a "point cloud image."

[0060] Step 3: The chest width estimation unit 202 creates an image by superimposing the infrared image and the point cloud image. Hereinafter, the image by superimposing the infrared image and the point cloud image will be referred to as a "superimposed image."

[0061] Step 4: The chest width estimation unit 202 calculates the number of pixels per 1 cm from the superimposed image. That is, the chest width estimation unit 202 calculates the number of pixels present between points in the y-axis direction (1 cm width). Hereinafter, the number of pixels present in a 1 cm width in the y-axis direction will be referred to as the "number of pixels per 1 cm."

[0062] Step 5: The chest width estimation unit 202 calculates the chest width from the number of pixels included in the width of the chest position in the area representing the breeding sow P that was the subject of the image capture. That is, the chest width estimation unit 202 calculates the chest width using the number of pixels that are on a line having an x-coordinate value that represents the chest position (this line is parallel to the y-axis) and that are included in the area representing the breeding sow P in the superimposed image, and the number of pixels per cm. The chest width is calculated as a / b, where a is the number of pixels that are on a line having an x-coordinate value that represents the chest position and that are included in the area representing the breeding sow P in the superimposed image, and b is the number of pixels per cm.

[0063] The x-coordinate value representing the chest position is set in advance. For example, as in Reference 1, the shoulder position of the breeding sow P is identified, and the position can be set to L cm above the shoulder position (i.e., toward the tail).

[0064] Furthermore, the area representing the breeding sow P in the superimposed image may be identified by any method. For example, an area corresponding to an area in the imaging guide 1100 may be identified as the area representing the breeding sow P, or, if Intel RealSense is used as the depth sensor (or depth camera) of the imaging device 10, each (x, y, z) coordinate value of the area of ​​the breeding sow P may be extracted and identified using a function called Threshold Filter provided in RealSense. In addition, the area representing the breeding sow P may be identified by performing image processing such as edge extraction or object area extraction on the infrared image or the superimposed image.

[0065] By performing the above steps 1 to 5, even if there is a defect in part of the point cloud data, the chest width of the breeding sow P that is the subject of the image capture can be estimated with high accuracy.

[0066] Next, the weight gain determination unit 203 determines the weight gain of the breeding sow P that was the subject of the image capture, using the number of days elapsed included in the image capture data received in step S201, the standard weight gain curve stored in the memory unit 206, and the chest width estimated in step S202 (step S203). An example of the standard weight gain curve is shown in FIG. 7. As shown in FIG. 7, the standard weight gain curve is expressed as a planar graph with the number of days elapsed since the first mating date on the horizontal axis and the standard chest width on the vertical axis. Hereinafter, the function representing the standard weight gain curve will be represented by f, and the standard chest width when the number of days elapsed is x will be represented by f(x).

[0067] When the number of days elapsed included in the photographed data received in the above step S201 is x' and the chest width estimated in the above step S202 is y', the weight gain determination unit 203 determines weight gain from the difference (deviation) between y' and f(x'). For example, the weight gain determination unit 203 determines weight gain as follows.

[0068] If |y'-f(x')|>th1 and y'-f(x')>0, the person is "too fat" If th1≧|y'-f(x')|>th2 and y'-f(x')>0, the person is overweight. If th2≧|y'-f(x')|≧0, it is "normal" If th1≧|y'-f(x')|>th2 and y'-f(x')<0, the person is underweight. If |y'-f(x')|>th1 and y'-f(x')<0, the person is "too thin" Here, th1 and th2 are preset thresholds, and th1>th2 is satisfied.

[0069] Next, the body weight gain determination result data creation unit 204 creates body weight gain determination result data including the body weight gain determination result determined in the above step S203 (step S204).

[0070] Then, the transmitting / receiving unit 201 transmits the body weight gain determination result data created in the above step S204 to the photographing device 10 that transmitted the photographing data received in the above step S201 (step S205). As a result, the body weight gain determination result determined in the above step S203 is displayed on the display of the photographing device 10.

[0071] <Standard weight gain curve generation process> The standard weight gain curve generation process according to this embodiment will be described below with reference to Fig. 8. In the following, it is assumed that the memory unit 206 of the weight gain determination device 20 stores a performance data set and a reference weight gain curve.

[0072] First, the standard body weight gain curve generating unit 205 acquires the standard body weight gain curve stored in the storage unit 206 (step S301). Hereinafter, the function representing the standard body weight gain curve is represented by g, and the chest width when the number of days elapsed is x is represented by g(x).

[0073] Next, the standard body weight gain curve generating unit 205 divides the number of days elapsed in the standard body weight gain curve acquired in the above step S301 into fixed intervals (for example, every 30 days) (step S302). Hereinafter, the k-th interval of the standard body weight gain curve will be referred to as I k (k=1, ,K) in the interval I k The standard weight gain curve at g k Here, K is the number of sections.

[0074] Next, the standard weight gain curve generating unit 205 adds a constant to the standard weight gain curve for each section to minimize the error with the actual data included in the actual data set stored in the memory unit 206, thereby creating a graph called an adjusted weight gain curve (step S303).

[0075] For example, the performance data set is {(x n ,y n )|n=1, ,N'}, where x n is the number of days elapsed in the nth actual data, y n is the chest width included in the nth performance data, and N' is the number of performance data included in the performance data set.

[0076] At this time, the standard weight gain curve generating unit 205 calculates the weight gain curve for section I. k For each x n ∈I k Error for |y n -(g k (x n )+C k )|| k Determine h k (x)=g k (x)+C k (x∈I k ) and then for k=1, ,K, the function h k (x)(x∈I k ) are obtained, so these functions h k (x)(x∈I k ) so x∈I k When h(x)=h k (x) and the function h(x)(x∈I1∪ ∪I K ) and the graph represented by this function h is the adjusted weight gain curve.

[0077] An example of the base weight gain curve and the adjusted weight gain curve is shown in Figure 9. Note that Figure 9 also shows points (actual data points) represented by each performance data. As shown in Figure 9, the adjusted weight gain curve is expressed by adding a certain constant to the base weight gain curve for each interval (30 days in the example shown in Figure 9).

[0078] Next, the standard weight gain curve generating unit 205 extracts the chest width from the adjusted weight gain curve for each day (step S304). That is, the standard weight gain curve generating unit 205 extracts (x, h(x)) for x = 0, 1, 2, . . . , N, where N is the maximum number of days that have passed. The value of N is set in advance.

[0079] Next, the standard weight gain curve generating unit 205 generates a standard weight gain curve by the least squares method using {(x, h(x))|x=0, 1, 2, . . . , N} extracted in step S304 (step S305). That is, the standard weight gain curve generating unit 205 calculates (h(x)-f(x)) for x=0, 1, 2, . . . , N).2 A function f that minimizes the sum of is found, and this function f is used as a function representing the standard weight gain curve. As a result, even if actual data including a chest width that is an outlier exists in the actual data set, it is possible to generate a robust standard weight gain curve that reduces the influence of the outlier. Note that the least squares method generally has the disadvantage of being easily influenced by outliers.

[0080] Then, the standard weight gain curve generating unit 205 stores the function f representing the standard weight gain curve generated in the above step S305 in the storage unit 206 (step S306). This enables the weight gain determination device 20 to determine the weight gain of the breeding sow P using the standard weight gain curve in the above weight gain determination process.

[0081] <Summary> As described above, the body weight gain determination system 1 according to this embodiment can accurately determine the body weight gain of a pig by the following (1) to (3).

[0082] (1) By using at least one of the above conditions 1 to 5 as the photographing condition, it becomes possible to automatically photograph the sow P when, for example, the relative positional relationship with the sow P is appropriate or the photographing environment is appropriate. This makes it possible to accurately estimate the chest width without being affected by, for example, the photographer's photographing technique, and as a result, to accurately determine weight gain.

[0083] (2) By performing the above steps 1 to 5, even if there are missing parts in the point cloud data, the chest width of the photographed breeding sow P can be estimated with high accuracy. As a result, it becomes possible to accurately determine the weight gain.

[0084] (3) The above-described standard weight gain curve generation process can generate a robust standard weight gain curve that reduces the influence of outliers, even if the actual data set contains outliers. As a result, it becomes possible to accurately determine weight gain.

[0085] <Modification> Variation 1 In the above embodiment, a photograph is automatically taken when predetermined photographing conditions are met, but it is not necessary to take a photograph automatically. For example, when predetermined photographing conditions are met, the user M may be notified of this and prompted to take a photograph.

[0086] Variation 2 In the above embodiment, the weight gain determination results are classified into five categories: "too thin," "slightly thin," "normal," "slightly overweight," and "too overweight." However, the present invention is not limited to this, and the weight gain determination results may be classified into less than five categories or six or more categories. Also, instead of determining weight gain, for example, weight may be estimated from chest width. Note that weight can be estimated from chest width using a known method (for example, a weight estimation model expressed by a regression equation, etc.).

[0087] Variation 3 In the above embodiment, three steps have been described: "automatic photography when photography conditions are met," "chest width estimation using steps 1 to 5 above," and "standard weight gain curve generation processing." However, the weight gain determination system 1 according to this embodiment is not limited to realizing all three of these steps, and may, for example, realize at least one of the three steps.

[0088] Variation 4 In the above embodiment, the performance data is created by, for example, having user M or the like actually measure the chest width of a breeding sow and associating it with the number of days elapsed since the breeding sow's first mating date, but the present invention is not limited to this. For example, the above-mentioned photographing process and chest width estimation in steps S201 to S202 of the above-mentioned weight gain determination process may be performed to create performance data including the number of days elapsed and chest width.

[0089] Variation 5 In the above embodiment, the subject of weight gain assessment was a breeding sow, but this is not limited to this and may be a general pig or livestock other than pigs (e.g., cows, horses, sheep, etc.).

[0090] The present invention is not limited to the above-described specifically disclosed embodiments, and various modifications, changes, and combinations with known technologies are possible without departing from the scope of the claims.

[0091] [References] Reference 1: JP 2022-109683 A [Explanation of symbols]

[0092] 1. Weight Gain Assessment System 10 Imaging equipment 20. Weight Gain Determination Device 101 Transmitter / Receiver 102 UI control section 103 Shooting condition determination unit 104 Photography Department 105 Shooting Data Creation Department 201 Transmitter / Receiver 202 Chest width estimation part 203 Body Gain Assessment Unit 204 Weight Gain Assessment Result Data Creation Department 205 Standard gain curve generator 206 Memory section

Claims

1. A body weight gain determination system including a photographing device that photographs a target livestock representing a livestock that is a target for body weight gain determination, and a body weight gain determination device that determines the body weight gain of the target livestock, The body weight gain determination device includes: a chest width estimation unit that estimates the chest width of the target livestock based on the photographing data acquired from the photographing device; a curve generating unit that generates a first curve representing the relationship between the number of days elapsed since the first mating date of the livestock and a standard chest width based on performance data that associates the number of days elapsed since the first mating date of the livestock with a chest width and a relationship between the number of days elapsed and a general chest width of the livestock at the number of days elapsed; and a weight gain determination unit that determines the weight gain of the livestock based on the chest width estimated by the chest width estimation unit and the first curve. Weight gain assessment system.

2. The photographing data includes the number of days elapsed, The weight gain determination unit The weight gain determination system described in claim 1, which determines the weight gain of the livestock from the magnitude of the difference between the chest width corresponding to the number of elapsed days included in the photographing data in the first curve and the chest width estimated by the chest width estimation unit.

3. The photographing data includes an infrared image of the livestock photographed by an infrared camera and point cloud data representing the results of measuring the depth of the livestock at each point within the photographing range by a depth sensor; The chest width estimation unit aligning the points included in the point cloud data in a width direction of the livestock by a predetermined unit width; calculating the number of pixels per unit width from a superimposed image representing an image obtained by superimposing an image of the aligned point cloud data and the infrared image; 3. The system for determining body weight gain according to claim 1, wherein the chest width is estimated based on the number of pixels included in the width of the chest position of the livestock and the number of pixels per unit width.

4. The curve generation unit A third curve is created by adding a constant to the chest width of the second curve, for each interval obtained by dividing the number of elapsed days of the second curve representing the relationship between the number of elapsed days and the general chest width of the livestock at the number of elapsed days into a certain interval, so that the error between the chest width of the second curve and the chest width of the actual data whose number of elapsed days falls within the interval is minimized; The system for determining body weight gain according to claim 1 , wherein the curve that minimizes the sum of squares of the difference between the chest width of the third curve and the chest width at each elapsed day is generated as the first curve.

5. a chest width estimation step of estimating the chest width of a target livestock based on photographic data acquired from a photographing device that photographs the target livestock representing the livestock whose body weight gain is to be determined; a curve generation step of generating a first curve representing the relationship between the number of days elapsed since the first mating date of the livestock and a standard chest width, based on performance data correlating the number of days elapsed since the first mating date of the livestock with the chest width, and the relationship between the number of days elapsed and a general chest width of the livestock at the number of days elapsed; a weight gain determination procedure for determining weight gain of the livestock based on the chest width estimated by the chest width estimation procedure and the first curve; A method for determining weight gain performed by a computer.

6. a chest width estimation step of estimating the chest width of a target livestock based on photographic data acquired from a photographing device that photographs the target livestock representing the livestock whose body weight gain is to be determined; a curve generation step of generating a first curve representing the relationship between the number of days elapsed since the first mating date of the livestock and a standard chest width, based on performance data correlating the number of days elapsed since the first mating date of the livestock with the chest width, and the relationship between the number of days elapsed and a general chest width of the livestock at the number of days elapsed; a weight gain determination procedure for determining weight gain of the livestock based on the chest width estimated by the chest width estimation procedure and the first curve; A program that causes a computer to execute the following.

Images