JP7648161B2 - Cattle evaluation device and cattle evaluation method - Google Patents

Description

The present invention relates to a cow evaluation device and a cow evaluation method.

To evaluate the energy balance in cattle feeding management, the state of body fat (fleshyness) of cattle is evaluated. One index showing the state of body fat is the body condition score. Normally, the body condition score is judged by veterinarians or dairy experts, but because it is judged by humans, the judgment criteria are vague and there is variation in the judgment. On the other hand, there are known methods for evaluating a score indicating the body condition based on a three-dimensional image of an animal (e.g., Patent Document 1) and a method for evaluating a score indicating the body condition based on a two-dimensional image of an animal (e.g., Patent Document 2).

Special Publication No. 2012-510278 European Patent No. 2027770

The evaluation method described in Patent Document 1 uses three-dimensional images of the animals to calculate a score indicating the physical condition of the animals, so it is necessary to take images at a specific location where large-scale equipment is installed, making it difficult to apply to a variety of breeding sites. The evaluation method described in Patent Document 2 uses two-dimensional images of the animals, but because it uses two-dimensional images taken from above, it is necessary to take images at a specific location where equipment capable of taking images from above is installed, and like Patent Document 1, it is difficult to apply to a variety of breeding sites.

The present invention has been made in consideration of the above problems, and aims to provide a cattle evaluation device and a cattle evaluation method that can be applied to a variety of breeding sites.

The cattle evaluation device of the present invention includes an image acquisition unit that acquires a two-dimensional image of a standing cattle from approximately directly behind or approximately directly in front, an extraction unit that extracts the contour line of the cattle from the captured image, a determination unit that identifies a reference dimension related to the size of the cattle's body in the width direction from the contour line, and identifies a reference range between two straight lines that are approximately perpendicular to each other in the direction away from the leftmost point on the left side, the rightmost point on the right side, or the topmost point on the top side of the contour line and are separated from the leftmost point on the left side, the rightmost point on the right side, or the topmost point on the top side of the contour line by a first distance corresponding to a first percentage of the reference dimension and a second distance corresponding to a second percentage of the reference dimension, and determines one or more straight lines from multiple points on a continuous portion of the contour line that is located within the reference range, a calculation unit that calculates the slope of the one or more straight lines determined by the determination unit or a value obtained from the slope, and an evaluation unit that evaluates the fleshiness of the cattle from the slope calculated by the calculation unit or the value obtained from the slope.

The present invention provides a cattle evaluation device and a cattle evaluation method that can be applied to a variety of breeding sites.

FIG. 1 is a block diagram showing the configuration of a cattle evaluation device according to the first embodiment. FIG. 2 is a block diagram showing the functional configuration of the control unit in the first embodiment. FIG. 3 is a diagram showing an example of an image acquired by the image acquisition unit. FIG. 4 is a diagram for explaining the process executed by the extraction unit. FIG. 5 is a diagram for explaining the processes executed by the determining unit and the calculating unit in the first embodiment. 6(a) to 6(c) are images of cows with different body sizes. 7(a) and 7(b) are diagrams showing the relationship between the fleshiness of a cow and the slope of a straight line calculated by the calculation unit. FIG. 8 shows the results of an experiment investigating the correlation between the body condition score and the maximum slope, minimum slope, and average slope of a plurality of straight lines calculated by the calculation unit. FIG. 9 is a block diagram showing the hardware configuration of the cow evaluation device according to the first embodiment. FIG. 10 is a flowchart showing an example of a method for evaluating beef meatiness in the first embodiment. FIG. 11 is a diagram for explaining the processes executed by the determining unit and the calculating unit in the modified example of the first embodiment. FIG. 12 is a diagram showing the results of an experiment investigating the correlation between the body condition score and the slope of the line calculated by the calculation unit. FIG. 13 is a diagram for explaining the processes executed by the determining unit and the calculating unit in the second embodiment. FIG. 14 is a diagram showing the results of an experiment investigating the correlation between the body condition score and the average slope of a plurality of straight lines calculated by the calculation unit. FIG. 15(a) is an image of a cow in an upright position captured from substantially directly behind, and FIG. 15(b) is an image of the cow in FIG. 15(a) captured from substantially the front. FIGS. 16(a) and 16(b) are images in which the contour lines of the cow are extracted from the captured images of FIGS. 15(a) and 15(b).

First Embodiment
A cow evaluation device 100 according to a first embodiment will be described below with reference to Figs. 1 to 10. Fig. 1 shows in a block diagram the configuration of the cow evaluation device 100 according to the first embodiment. Fig. 2 shows in a block diagram the functional configuration of a control unit 22. In this first embodiment, the evaluation device 100 is a portable information device carried and used by an operator. Examples of portable information devices include smartphones, tablet computers, PDAs (Personal Digital Assistants), and smart glasses. The evaluation device 100 has, for example, a telephone function, an email function, a communication function for connecting to the Internet, and a data processing function for executing a program.

As shown in FIG. 1, the evaluation device 100 includes a display unit 12, an operation unit 14, a communication unit 16, an imaging unit 18, a memory unit 20, and a control unit 22.

The display unit 12 is a display device that displays images, various information, and images for operation input such as touch operation buttons.

The operation unit 14 is a device such as a touch panel or a switch. The touch panel accepts information input in response to a worker's touch and transmits the accepted information to the control unit 22. The touch panel is, for example, incorporated in the display unit 12. Thus, the touch panel accepts various information inputs in response to the worker touching the surface of the display unit 12. The switch is an operating member that accepts operations on the evaluation device 100 from the worker and transmits the accepted information to the control unit 22.

The communication unit 16 is a communication interface that performs short-range wireless communication (e.g., NFC (Near Field Communication)) with other devices, and wireless communication (e.g., communication using a mobile phone line or a wireless LAN (Local Area Network)) with other devices connected to the network.

The imaging unit 18 is a device that captures images such as still images and videos. The imaging unit 18 captures images in response to an operator carrying the evaluation device 100 operating the operation unit 14. When evaluating the fleshiness of a cow, the operator uses the imaging unit 18 to capture an image of the cow in a standing position from approximately directly behind. Capturing an image from approximately directly behind means capturing an image in a range that captures the outline of the cow's left and right hip angles. Therefore, the operator captures an image from a position several meters (e.g., 1 m to 2 m) behind the cow so that the outline of the cow's left and right hip angles is captured.

The memory unit 20 stores various information such as captured images of the cow, extracted images of the cow's contours, and evaluation results of the cow's body fatness.

The control unit 22 is a processing unit that controls each part of the cow evaluation device 100. As shown in FIG. 2, the control unit 22 has an image acquisition unit 30, an extraction unit 32, a determination unit 38, a calculation unit 34, and an evaluation unit 36.

Figure 3 shows an example of an image acquired by the image acquisition unit 30. As shown in Figure 3, the image acquisition unit 30 acquires an image 50, for example, of a cow in an upright position, captured from approximately directly behind. The image 50 of the cow to be acquired may be an image captured by the imaging unit 18, or an image captured by another device and received by the communication unit 16. The image 50 of the cow to be acquired may be an image selected by the operator using the operation unit 14 from those stored in the memory unit 20. The image acquisition unit 30 may display the acquired image 50 of the cow on the display unit 12. This allows the operator to visually recognize on the display unit 12 the cow to be evaluated for its fleshiness.

Figure 4 is a diagram for explaining the processing executed by the extraction unit 32. As shown in Figure 4, the extraction unit 32 extracts a contour line 60 that traces the outline of the cow from the captured image 50 of the cow acquired by the image acquisition unit 30. A commonly known image processing technique can be used to extract the contour line 60. The extraction unit 32 may display an extracted image 51 that shows only the extracted contour line 60 of the cow on the display unit 12. This allows the operator to visually recognize on the display unit 12 the outline of the cow whose fleshiness is to be evaluated. The outer shape of the extracted image 51 is the same as that of the captured image 50, and the position of the cow relative to the image frame is the same in the extracted image 51 and the captured image 50.

There are peaks 61a and 61b on both the left and right sides of the upper part of the contour line 60. The peaks 61a and 61b correspond to the cow's hips. The peak 61a has a vertex 62a at the approximate center of the peak 61a on the contour line 60, a first curved portion 63a extending from the vertex 62a to the upper center side of the contour line 60 (the cow's spine side), and a second curved portion 64a extending from the vertex 62a to the lower side of the side of the contour line 60 (the cow's belly side). Similarly, the peak 61b has a vertex 62b at the approximate center of the peak 61b on the contour line 60, a first curved portion 63b extending from the vertex 62b to the upper center side of the contour line 60 (the cow's spine side), and a second curved portion 64b extending from the vertex 62b to the lower side of the contour line 60 (the cow's belly side). The first curved portions 63a, 63b can be considered to be curved portions that extend from approximately the center point (vertices 62a, 62b) on the contour line 60 of the peaks 61a, 61b to the other peak side (the cow's spine side). The second curved portions 64a, 64b can be considered to be curved portions that extend from approximately the center point (vertices 62a, 62b) on the contour line 60 of the peaks 61a, 61b to the opposite side to the first curved portions 63a, 63b.

5 is a diagram for explaining the processing executed by the determination unit 38 and the calculation unit 34 in the first embodiment. FIG. 5 is an enlarged view of the upper part of the contour line 60, and is illustrated diagrammatically for clarity (the same applies to the following similar figures). As shown in FIG. 5, the determination unit 38 first identifies a reference dimension related to the width direction size of the cow's body from the contour line 60. In this first embodiment, the distance W between a vertical line 304 that passes through the leftmost point 302 on the left side of the contour line 60 and extends vertically (up and down) in the extracted image 51, and a vertical line 305 that passes through the rightmost point 303 on the right side of the contour line 60 and extends vertically (up and down) in the extracted image 51 is identified as the reference dimension. For example, the leftmost point 302 can be determined by moving a vertical line extending in the up and down direction in the extracted image 51 closer to the contour line 60 from the left side, and identifying the point where the vertical line first touches the contour line 60 as the leftmost point 302. Alternatively, the leftmost point 302 may be identified by other methods. For example, the extracted image 51 may be converted into an X-Y coordinate system and the leftmost point 302 may be identified from the X and Y coordinates of each pixel of the extracted image 51, or the leftmost point 302 may be determined to be the point on the contour 60 that is the shortest distance from the left edge of the extracted image 51. The same applies to the rightmost point 303.

Next, the determination unit 38 determines a first distance corresponding to a first ratio of the determined reference dimension W, and a second distance corresponding to a second ratio of the reference dimension W. The determination unit 38 then determines a reference range 306 sandwiched between a straight line 330 that is separated from the leftmost point 302 of the contour line 60 by the first distance inside the contour line 60 and is substantially perpendicular (for example, perpendicular) in the direction away from the leftmost point 302, and a straight line 332 that is separated from the leftmost point 302 by the second distance inside the contour line 60 and is substantially perpendicular (for example, perpendicular) in the direction away from the leftmost point 302. In this first embodiment, the determination unit 38 uses 20% as the first ratio to determine 0.2W as the first distance, and uses 30% as the second ratio to determine 0.3W as the second distance. Therefore, it can also be said that the reference range 306 is a range that is separated from the vertical line 304 passing through the leftmost point 302 by 0.2W or more and 0.3W or less. Next, the determination unit 38 identifies a plurality of points 307 on a continuous portion of the contour 60 that is located within the reference range 306, and determines a plurality of straight lines 310a, 310b, 310c, 310d, and 310e that pass through two of the plurality of points 307. In this first embodiment, the plurality of points 307 are two points where the straight lines 330 and 332 intersect with the contour 60, and four points where the contour 60 intersects with four vertical lines that divide the reference range 306 into five equal parts in the left-right direction. In this first embodiment, the plurality of straight lines 310a to 310e are five straight lines 310a to 310e that pass through each of the five pairs of two adjacent points obtained from the six points 307.

The calculation unit 34 calculates the inclination of the straight lines 310a-310e determined by the determination unit 38 with respect to a horizontal line extending horizontally (left-right) in the extracted image 51. In this first embodiment, the inclination of the straight lines 310a-310e determined by the determination unit 38 is positive when they extend from the bottom left to the top right in the extracted image 51, and negative when they extend from the bottom right to the top left. The calculation unit 34 determines the maximum inclination, minimum inclination, or average inclination from the inclination of each of the multiple straight lines 310a-310e. The inclination of the straight lines 310a-310e calculated by the calculation unit 34 is one of the index values indicating the degree of inclination of the peak portion 61a.

The evaluation unit 36 evaluates the fleshiness of the cow captured in the captured image 50 from the maximum slope, minimum slope, or average slope of the lines 310a to 310e calculated by the calculation unit 34. For example, the body condition score is evaluated as the fleshiness.

Here, we will explain the relationship between the maximum slope, minimum slope, or average slope of the lines 310a to 310e calculated by the calculation unit 34 and the body condition score. The body condition score is an index that indicates the state of body fat deposition, and may be assigned a score of, for example, 1.00 to 5.00, with the higher the value, the more fleshy the body is.

Figures 6(a) to 6(c) are images 50 of cows with different levels of fleshiness. Figure 6(a) is an image 50 of a cow with poor fleshiness, Figure 6(b) is an image 50 of a cow that is more fleshy than the cow in Figure 6(a), and Figure 6(c) is an image 50 of a cow that is even more fleshy. The cow in Figure 6(a) was given a body condition score of 2.25 by an expert. The cow in Figure 6(b) was given a score of 3.00, and the cow in Figure 6(c) was given a score of 4.00. As shown in Figures 6(a) to 6(c), it can be seen that the fleshiness around the hip angle changes due to differences in fleshiness (i.e., different body condition scores), and the shape of the hip angle changes when the cow is viewed from directly behind in a standing position.

Figures 7(a) and 7(b) are diagrams showing the relationship between the fleshiness of a cow and the slope of multiple straight lines 310a-310e calculated by calculation unit 34. Figure 7(b) shows the contour line 60 of a cow that is fleshier than the cow in Figure 7(a). As shown in Figures 7(a) and 7(b), the slope of the straight lines 310a-310e varies depending on the fleshiness of the cow. From this, it is thought that the fleshiness of a cow can be evaluated by determining the slope of the straight lines 310a-310e.

[experiment]
An experiment was conducted to investigate the correlation between the body condition score evaluated by a veterinarian or other expert and the maximum slope, minimum slope, and average slope of the straight lines 310a to 310e calculated by the calculation unit 34. The experiment was conducted on dairy cows (Holstein breed).

FIG. 8 shows the results of an experiment investigating the correlation between the body condition score and the maximum slope, minimum slope, and average slope of the multiple straight lines 310a to 310e calculated by the calculation unit 34. The horizontal axis (x-axis) of FIG. 8 is the body condition score evaluated by an expert, and the vertical axis is the slope calculated by the calculation unit 34. The multiple black circles in FIG. 8 show the results for each dairy cow when the maximum slope of the straight lines 310a to 310e is used, and the dotted straight line shows the approximation straight line (regression line) at that time. The multiple black squares show the results for each dairy cow when the minimum slope of the straight lines 310a to 310e is used, and the dashed straight line shows the approximation straight line (regression line) at that time. The multiple black triangles show the results for each dairy cow when the average slope of the straight lines 310a to 310e is used, and the dashed straight line shows the approximation straight line (regression line) at that time. As shown in FIG. 8, regardless of whether the maximum slope, minimum slope, or average slope of the straight lines 310a to 310e is used, the body condition score decreases as the negative value of the slope increases, and the body condition score increases as the positive value of the slope increases. The approximate dotted line using the maximum slope was calculated as y = 0.2084x - 0.4895, and the coefficient of determination ^R2 was 0.6317. The approximate dotted line using the minimum slope was calculated as y = 0.2528x - 1.0665, and the coefficient of determination ^R2 was 0.8772. The approximate dashed line using the average slope was calculated as y = 0.2203x - 0.7692, and the coefficient of determination ^R2 was 0.8996. This confirmed that there is a strong correlation between the body condition score and the slope of the straight lines 310a to 310e calculated by the calculation unit 34.

As described above, since there is a strong correlation between the body condition score and the slope of the lines 310a-310e calculated by the calculation unit 34, it can be seen that the evaluation unit 36 can accurately evaluate the beefiness of the cow by evaluating the maximum slope, minimum slope, or average slope of the lines 310a-310e. The evaluation unit 36 evaluates the body condition score using, for example, score information stored in the memory unit 20 that associates the slope with the body condition score.

Table 1 shows an example of score information stored in the storage unit 20. As shown in Table 1, in the score information, the slope and the body condition score are associated with each other. For example, when the slope is less than -a, the body condition score is 1.00, when the slope is equal to or greater than -a and less than -b, the body condition score is 1.50, ..., when the slope is equal to or greater than -d and less than +e, the body condition score is 3.00, and so on. Note that, although Table 1 shows an example in which the body condition score is in increments of 0.5, other cases are also possible. For example, the body condition score may be in increments of 0.25, or between 2.00 and 4.00, the increments may be 0.25, and the rest may be in increments of 0.50. Note that separate tables are used for the maximum slope, the minimum slope, and the average slope.

The score information stored in the memory unit 20 may be other than the score information shown in Table 1, as long as it is information that associates the slope with the body condition score. For example, a linear function (y = ax + b) such as the approximated straight line shown in Figure 8 may be stored in the memory unit 20 as the score information. In this case, the body condition score may be evaluated by performing fractional processing such as rounding up, rounding up, or rounding down on the value of x obtained by substituting the slope for y.

Figure 9 is a block diagram showing the hardware configuration of the cow evaluation device 100 according to the first embodiment. As shown in Figure 9, the cow evaluation device 100 has a storage device 70, a memory 71, a processor 72, a communication interface 73, an input device 74, a display device 75, and an imaging device 76. These components are interconnected by a bus 77.

The storage device 70 is a non-volatile semiconductor memory such as a flash memory, and stores the cow evaluation program 78.

Memory 71 is hardware that temporarily stores data, such as a dynamic random access memory (DRAM). The memory unit 20 in FIG. 1 is realized by the storage device 70 and memory 71.

The processor 72 is hardware such as a CPU (Central Processing Unit) that controls each part of the cow evaluation device 100. The processor 72 executes the cow evaluation program 78 in cooperation with the memory 71. In this way, the memory 71 and the processor 72 cooperate to execute the cow evaluation program 78, thereby realizing each function of the control unit 22 in FIG. 2.

The communication interface 73 is hardware that enables the cow evaluation device 100 to communicate with other devices through short-range wireless communication or wireless communication via a network. The communication interface 73 realizes the communication unit 16 in FIG. 1.

The input device 74 is an input device such as a touch panel or a switch that allows the operator to operate the cattle evaluation device 100. The input device 74 realizes the operation unit 14 in FIG. 1.

The display device 75 is hardware for implementing the display unit 12 in FIG. 1, and is a display device such as an LCD display for displaying the captured image 50 of the cow, the extracted image 51, the evaluation results of the beefiness of the cow, etc.

The imaging device 76 is hardware for implementing the imaging unit 18 in FIG. 1, and includes a lens, an imaging element, etc.

Next, an example of a method for evaluating beef meatiness in the first embodiment will be described with reference to the flowchart in FIG. 10. The process in FIG. 10 is executed by the control unit 22 of the evaluation device 100. It is assumed that the memory unit 20 stores score information such as that shown in Table 1, for example.

In the process of FIG. 10, first, in step S10, the image acquisition unit 30 of the control unit 22 waits until the cow evaluation program 78 is started. In this case, the process moves to step S12 when the control unit 22 starts the cow evaluation program 78 in response to an operation by the operator (e.g., an input operation to the operation unit 14 or a voice input to a microphone (not shown)).

When the process proceeds to step S12, the image acquisition unit 30 acquires an image 50 of the cow to be evaluated for meatiness (see FIG. 3). The image 50 of the cow is an image of the cow in an upright position captured from almost directly behind. The image acquisition unit 30 acquires, for example, an image of the cow captured by the operator using the imaging unit 18. The image acquisition unit 30 may acquire an image of the cow received from another device via the communication unit 16, or an image of the cow selected by the operator by operating the operation unit 14 from among multiple images of the cow stored in the storage unit 20. The image acquisition unit 30 may start a camera function so that the operator can capture an image of the cow using the imaging unit 18 after the cow evaluation program 78 is started. When the operator captures an image of the cow using the imaging unit 18, the image of the captured cow and a selection button asking whether or not to select this image may be displayed on the display unit 12, and if the operator chooses not to select the image, the camera function may be started again.

In step S14, the extraction unit 32 of the control unit 22 extracts a contour line 60 that traces the outline of the cow from the captured image 50 of the cow acquired in step S12 (see FIG. 4). This generates an extracted image 51 showing the cow's contour line 60. When extracting the contour line 60, the extraction unit 32 may perform a process to remove background other than the cow from the captured image 50 in order to extract the contour line 60 with good accuracy.

Next, in step S16, the determination unit 38 of the control unit 22 specifies a reference dimension for the width direction of the cow's body from the contour line 60, and specifies a reference range 306 between a straight line 330 that is a first distance away from the leftmost point 302 of the contour line 60, which corresponds to a first percentage of the reference dimension, and a straight line 332 that is a second distance away from the leftmost point 302 of the contour line 60, which corresponds to a second percentage of the reference dimension (see FIG. 5). Then, the determination unit 38 specifies multiple straight lines 310a to 310e from multiple points 307 on a continuous portion of the contour line 60 that is located within the reference range 306 (see FIG. 5). In this first embodiment, the determination unit 38 specifies the distance W between the vertical line 304 that passes through the leftmost point 302 of the contour line 60 and the vertical line 305 that passes through the rightmost point 303 of the contour line 60 as the reference dimension for the width direction of the cow's body. The determination unit 38 then identifies a reference range 306 between a line 330 that is 0.2W away from the leftmost point 302 of the contour 60 and a line 332 that is 0.3W away from the leftmost point 302 of the contour 60, and determines a number of straight lines 310a to 310e that pass through pairs of two adjacent points among a number of points 307 on a continuous portion of the contour 60 that is located within the reference range 306 (see FIG. 5).

Next, in step S18, the calculation unit 34 of the control unit 22 calculates the maximum slope, minimum slope, or average slope of the multiple straight lines 310a to 310e determined in step S16.

Next, in step S20, the evaluation unit 36 of the control unit 22 evaluates the body condition score of the cow captured in the captured image 50 based on the maximum slope, minimum slope, or average slope of the lines 310a-310e calculated in step S18 and the score information stored in the memory unit 20, such as that shown in Table 1.

Next, in step S22, the evaluation unit 36 displays the body condition score evaluated in step S20 on the display unit 12 and stores it in the memory unit 20. For example, the evaluation unit 36 may display the body condition score on the display unit 12 together with the captured image 50 of the cow. The evaluation unit 36 may also associate individual identification information, such as the individual identification number of the cow, input by the operator operating the operation unit 14 with the evaluated body condition score and store it in the memory unit 20. Note that the evaluation unit 36 is not limited to acquiring individual identification information input by the operator operating the operation unit 14, but may also acquire individual identification information from an image captured by the imaging unit 18 of an image of an identification tag attached to the body of the cow, or may acquire individual identification information by using RFID (Radio Frequency Identification) technology and having the communication unit 16 receive the individual identification information from an RF tag attached to the body of the cow. Alternatively, cameras and communication devices may be installed in the passages inside and outside the barn so that the sensors can detect cows passing through the passages, and the images of the cows passing through the passages captured by the cameras may be acquired by the communication unit 16, and the individual identification information may be acquired from the images of the identification tags captured in the images. Alternatively, the communication device may receive the individual identification information from RF tags attached to the cows passing through the passages, and the communication unit 16 may receive the individual identification information from the communication device. Also, the evaluation unit 36 may store in the memory unit 20 the individual identification number of the cow whose body condition score has been evaluated, the image 50 of the cow, the image date, and the body condition score of the cow as a single piece of associated information.

[Modification]
In the above first embodiment, the determination unit 38 determines a plurality of straight lines 310a to 310e, but a single straight line may be determined. Fig. 11 is a diagram for explaining the process executed by the determination unit 38 and the calculation unit 34 in the modified example of the first embodiment. As shown in Fig. 11, the determination unit 38 specifies the reference range 306 by the same method as that described in Fig. 5. Then, the determination unit 38 specifies points 307 on the contour line 60 at both ends of the reference range 306, and determines the straight line 310 passing through the two points 307.

The calculation unit 34 calculates the slope of the straight line 310 determined by the determination unit 38. The evaluation unit 36 evaluates the beefiness of the cow from the slope calculated by the calculation unit 34.

[experiment]
An experiment was conducted to investigate the correlation between the body condition score evaluated by a veterinarian or other expert and the gradient of the straight line 310 calculated by the calculation unit 34. The experiment was conducted on dairy cows (Holstein breed).

FIG. 12 shows the results of an experiment investigating the correlation between the body condition score and the slope of the straight line 310 calculated by the calculation unit 34. The horizontal axis (x-axis) of FIG. 12 is the body condition score evaluated by an expert, and the vertical axis is the slope calculated by the calculation unit 34. A plurality of black circles in FIG. 12 indicate the results of each dairy cow in the experiment, and the dotted straight line indicates an approximation line (regression line). As shown in FIG. 12, even when the slope of the straight line 310 passing through two points 307 on the contour line 60 at both ends of the reference range 306 is used, the body condition score becomes lower as the negative value of the slope increases, and the body condition score becomes higher as the positive value of the slope increases. The approximation line was found to be y = 0.2223x - 0.7697, and the coefficient of determination ^R2 was 0.8844. From this, it can be seen that the beefiness of a cow can be evaluated with high accuracy even when the beefiness is evaluated from the slope of a straight line 310 passing through two points 307 on the contour line 60 at both ends of the reference range 306.

As described above, according to the first embodiment and its modified examples, the image acquisition unit 30 acquires a two-dimensional captured image 50 of a cow in an upright position captured from approximately directly behind (see FIG. 3). The extraction unit 32 extracts the cow's contour line 60 from the captured image 50 (see FIG. 4). The determination unit 38 identifies the distance W between the vertical line 304 passing through the leftmost point 302 of the contour line 60 and the vertical line 305 passing through the rightmost point 303 of the contour line 60 as a reference dimension indicating the width of the cow's body from the contour line 60 (see FIG. 5). Then, the determination unit 38 specifies a reference range 306 between a straight line 330 located inside the contour line 60 from the leftmost point 302 on the contour line 60 by a first distance corresponding to 20% (first ratio) of the reference dimension W, and a straight line 332 located inside the contour line 60 by a second distance corresponding to 30% (second ratio), and determines one or more straight lines 310 or 310a to 310e from a plurality of points 307 on a continuous portion of the contour line 60 located within the reference range 306 (see FIG. 5 and FIG. 11). The calculation unit 34 calculates the slope of the straight line 310 or 310a to 310e determined by the determination unit 38 or an average slope obtained from each slope. The evaluation unit 36 evaluates the fleshiness of the cow captured in the captured image 50 from the slope calculated by the calculation unit 34 or the average slope obtained from the slopes. In this way, in the first embodiment and its modified example, the beefiness of the cow is evaluated using the two-dimensional captured image 50 captured from approximately behind the cow in an upright position, so that special equipment for capturing images of the cow in three dimensions or for capturing images of the cow from above is not required, and the method can be applied to a variety of breeding sites. Furthermore, even if the cow is captured large or small in the captured image 50, which is a two-dimensional image, the slope of the line does not change, so that the beefiness of the cow can be evaluated from the slope of the line, and the beefiness of the cow can be evaluated well without performing processing such as normalization. Furthermore, when a person evaluates the beefiness of the cow, the judgment criteria become ambiguous and the judgment varies, but in the first embodiment and its modified example, the control unit 22 judges the beefiness of the cow, so the judgment variation can be suppressed. Furthermore, since the beefiness of the cow is evaluated using the two-dimensional captured image 50 captured from approximately directly behind the cow in an upright position, the evaluation device 100, which is highly portable and highly versatile, such as a smartphone, can be used.

In the first embodiment and its modified examples, as shown in Figs. 5 and 11, the determination unit 38 determines the distance W between the vertical line 304 passing through the leftmost point 302 of the contour line 60 and the vertical line 305 passing through the rightmost point 303 of the contour line 60 as the reference dimension in the width direction of the cow's body. A reference range 306 is determined using this reference dimension W, and the slope of a straight line 310 or 310a-310e determined from multiple points 307 on the contour line 60 within the determined reference range 306 has a high correlation with the body condition score, as shown in Figs. 8 and 12. Therefore, by using this reference dimension W, the meatiness of the cow can be evaluated with high accuracy.

In a modified version of the first embodiment, as shown in FIG. 11, the determination unit 38 determines a straight line 310 that passes through two points 307 on the contour line 60 at both ends of the reference range 306. Determining the straight line 310 can be done with simpler processing than determining the multiple straight lines 310a-310e, and the slope of the straight line 310 can be calculated with simpler processing than calculating the slopes of the straight lines 310a-310e. Even when the slope of the straight line 310 is used, there is a high correlation with the body condition score, as shown in FIG. 12. Therefore, the fleshiness of the cow can be accurately evaluated with simple processing.

In the first embodiment, when the determination unit 38 determines a plurality of straight lines 310a-310e, it is preferable that the calculation unit 34 calculates the average slope of the plurality of straight lines 310a-310e. As described in Fig. 8, the average slope has a high coefficient of determination ^R2 , which indicates a correlation with the body condition score. Therefore, by the evaluation unit 36 evaluating the body condition score from the average slope, the meatiness of the cattle can be evaluated with high accuracy.

[Regarding reference range 306]
In the above first embodiment and its modified example, a case has been shown in which the range sandwiched between a straight line 330 that is a first distance away from the leftmost point 302 of the contour line 60 inside the contour line 60, the first distance corresponding to 20% (first ratio) of the reference dimension W, and a straight line 332 that is a second distance away from the leftmost point 302 of the contour line 60, the second distance corresponding to 30% (second ratio) of the reference dimension W, is set as the reference range 306. Here, a description will be given of how far the range from the leftmost point 302 of the contour line 60 should preferably be set as the reference range 306.

[experiment]
A range from the leftmost point 302 of the contour line 60 to the inside of the contour line 60 at a distance of at least a first percentage and at most a second percentage of the reference dimension W was defined as a reference range 306, and a plurality of reference ranges 306 were specified by changing the first percentage and the second percentage to various values. Then, as in Figures 8 and 12, a correlation was found between the body condition score evaluated by a veterinarian or other expert and a slope passing through two of the plurality of points 307 on the contour line 60 in each of the plurality of reference ranges 306, and a coefficient of determination ^R2 was calculated from an approximation line showing the correlation.

Tables 2 and 3 show the coefficients of determination R2 obtained from the approximation line showing the correlation between the body condition score and the slope of the line for each of a plurality of reference ranges 306 with different distances from the leftmost point 302 of the contour line 60. The reference ranges 306 and the coefficients of determination ^R2 on the left side (lower side) of Table ² were obtained from the approximation line showing the correlation between the body condition score and the slope of a line passing through two points on the contour line 60 at both ends of the reference range 306. The reference ranges 306 and the coefficients of determination ^R2 on the center and right side (upper side) were obtained from the approximation line showing the correlation between the body condition score and the average slope of a plurality of lines passing through two adjacent points among the plurality of points 307 on the contour line 60 located within the reference range 306 at an interval of 0.05W. Table 3 was obtained from the approximation line showing the correlation between the body condition score and the average slope of a plurality of lines passing through two adjacent points among the plurality of points 307 on the contour line 60 located within the reference range 306 at an interval of 0.02W.

As shown in Tables 2 and 3, a large coefficient of determination ^R2 was obtained when the reference range was set to 0.2 W to 0.3 W, 0.2 W to 0.35 W, 0.2 W to 0.4 W, 0.16 W to 0.34 W, 0.18 W to 0.32 W, 0.22 W to 0.3 W, and 0.22 W to 0.28 W. From this, it can be seen that in the case where the reference range 306 is a range away from the leftmost point 302 of the contour line 60 by a first percentage or more and a second percentage or less of the reference dimension W, a large coefficient of determination ^R2 can be obtained by setting the first percentage to a value of 16% or more and the second percentage to a value of 40% or less.

Therefore, when the determination unit 38 specifies the reference range 306 between a straight line 330 that is a first distance away from the leftmost point 302 of the contour line 60 inside the contour line 60, which corresponds to a first ratio of the reference dimension W, and a straight line 332 that is a second distance away from the leftmost point 302 of the contour line 60, which corresponds to a second ratio of the reference dimension W, it is preferable to use a value of 16% or more and a value of 40% or less as the second ratio. This allows the beefiness of the cow to be evaluated with high accuracy. In terms of evaluating the beefiness of the cow with high accuracy, it is preferable to use a value of 16% or more and 24% or less as the first ratio, more preferably a value of 18% or more and 22% or less, and even more preferably a value of 19% or more and 21% or less. It is preferable to use a value of 28% or more and 40% or less as the second ratio, more preferably a value of 30% or more and 38% or less, and even more preferably a value of 34% or more and 36% or less. Additionally, the width of the reference range 306 in the horizontal direction (the cow's width direction) within the extracted image 51 is preferably 0.05W or more, more preferably 0.08W or more, and even more preferably 0.1W or more.

In the above-mentioned first embodiment and its modified example, the determination unit 38 specifies the range between the straight line 330, which is a first distance away from the leftmost point 302 of the contour line 60 inside the contour line 60, corresponding to the first ratio of the reference dimension W, and the straight line 332, which is a second distance away from the leftmost point 302 of the contour line 60, corresponding to the second ratio of the reference dimension W, as the reference range 306. However, the present invention is not limited to this case, and the determination unit 38 may specify the range between the straight line, which is a first distance away from the rightmost point 303 of the contour line 60 inside the contour line 60, corresponding to the first ratio of the reference dimension W, and the straight line, which is a second distance away from the rightmost point 303 of the contour line 60, corresponding to the second ratio of the reference dimension W, as the reference range 306. Even in this case, it is preferable to use the above-mentioned values for the first ratio and the second ratio. When the rightmost point 303 is used to specify the reference range 306, the relationship between the positive/negative slope of the straight line passing through the multiple points 307 on the contour line 60 within the reference range 306 and the body condition score is opposite to that when the leftmost point 302 is used to specify the reference range 306. In other words, the more positive the slope of the line, the lower the body condition score, and the more negative the slope, the higher the body condition score.

In the first embodiment, the calculation unit 34 may calculate the median of the slopes of the multiple straight lines 310a to 310e that pass through the multiple points 307 on the contour line 60 within the reference range 306.

In the first embodiment, the determination unit 38 may determine a straight line passing through a pair of two points (two points every other) located on either side of one of the multiple points 307 on the contour line 60 within the reference range 306, or a straight line passing through a pair of two points (two points every other) located on either side of two points. The determination unit 38 may also determine an approximation line (regression line) obtained from the multiple points 307 by the least squares method.

In the first embodiment, the evaluation unit 36 may evaluate the beefiness of the cow based on the maximum value, minimum value, or average value of the multiple body condition scores obtained from the slopes of the multiple straight lines 310a-310e passing through the multiple points 307 on the contour line 60 within the reference range 306. The evaluation unit 36 may also evaluate the beefiness of the cow based on the average value of the body condition score obtained from the slope of one or more straight lines determined in the reference range 306 obtained from the leftmost point 302 and the body condition score obtained from the slope of one or more straight lines determined in the reference range 306 obtained from the rightmost point 303.

Second Embodiment
In the above first embodiment and its modified example, a case has been shown in which the determination unit 38 determines a range sandwiched between a straight line that is separated from the leftmost point 302 or the rightmost point 303 of the contour 60 by a first distance corresponding to a first proportion of the reference dimension inside the contour 60 and a straight line that is separated by a second distance corresponding to a second proportion of the reference dimension as the reference range 306. In the present second embodiment, a case will be described in which the determination unit 38 determines a range sandwiched between a straight line that is separated from the uppermost point 91 of the topmost side of the contour 60 by a first distance corresponding to the first proportion of the reference dimension inside the contour 60 and a straight line that is separated by a second distance corresponding to the second proportion of the reference dimension as the reference range 306a.

13 is a diagram for explaining the process executed by the determination unit 38 and the calculation unit 34 in the second embodiment. As shown in FIG. 13, the determination unit 38, like the first embodiment, specifies the distance W between the vertical line 304 passing through the leftmost point 302 of the contour line 60 and the vertical line 305 passing through the rightmost point 303 of the contour line 60 as the reference dimension. Next, the determination unit 38 specifies the range sandwiched between a straight line 330a that is separated from the top point 91 of the contour line 60 by a first distance (0.2W) corresponding to 20% (first ratio) of the reference dimension W inside the contour line 60 and is approximately perpendicular to the separated direction, and a straight line 332a that is separated by a second distance (0.3W) corresponding to 30% (second ratio) of the reference dimension W and is approximately perpendicular to the separated direction, as the reference range 306a. In other words, it can be said that the reference range 306a is a range that is separated from the horizontal line passing through the top point 91 by 0.2W or more and 0.3W or less. The top point 91 can be determined, for example, by moving a horizontal line in the extracted image 51 closer to the contour line 60 from above, and identifying the point where the horizontal line first touches the contour line 60 as the top point 91. Alternatively, the top point 91 may be determined by other methods, for example by converting the extracted image 51 into an X-Y coordinate system and identifying the top point 91 from the X and Y coordinates of each pixel of the extracted image 51, or by identifying the point on the contour line 60 that is the closest to the top edge of the extracted image 51.

Then, the determination unit 38 identifies a plurality of points 307a on a continuous portion of the contour 60 that is located within the reference range 306a, and determines a plurality of straight lines 312a, 312b that pass through two of the plurality of points 307a. In this second embodiment, a plurality of points 307a that are located within the reference range 306a in the left portion of the contour 60 are identified. The plurality of points 307a are two points where the straight lines 330a, 332a intersect with the contour 60, and one point where the contour 60 intersects with one horizontal line that divides the reference range 306a in two equal parts in the vertical direction. In this second embodiment, the plurality of straight lines 312a, 312b are two straight lines 312a, 312b that pass through each of two sets of adjacent two-point combinations obtained from the three points 307a.

The calculation unit 34 calculates the inclination of the straight lines 312a and 312b determined by the determination unit 38. In this second embodiment, the extraction image 51 is rotated 90° clockwise (i.e., the top point 91 of the contour line 60 is positioned on the right side), and the inclination of the straight lines 312a and 312b with respect to a horizontal line extending in the horizontal direction (left-right direction) in the rotational extraction image 51 is calculated. In this second embodiment, the inclination of the straight line is positive when it extends from the lower right to the upper left in the rotational extraction image 51, and negative when it extends from the lower left to the upper right. The calculation unit 34 calculates the maximum inclination, minimum inclination, or average inclination from the inclination of the straight lines 312a and 312b, but in this second embodiment, the average inclination is calculated. The inclination of the straight lines 312a and 312b calculated by the calculation unit 34 is one of the index values indicating the degree of inclination of the peak portion 61a.

The evaluation unit 36 evaluates the fleshiness of the cow captured in the captured image 50 from the average slope of the lines 312a and 310b calculated by the calculation unit 34. As the fleshiness, for example, the body condition score is evaluated.

[experiment]
An experiment was conducted to investigate the correlation between the body condition score evaluated by a veterinarian or other expert and the average slope of the multiple straight lines 312a, 312b calculated by the calculation unit 34. The experiment was conducted on dairy cows (Holstein breed).

FIG. 14 shows the results of an experiment investigating the correlation between the body condition score and the average slope of the multiple straight lines 312a and 312b calculated by the calculation unit 34. The horizontal axis (x-axis) of FIG. 14 is the body condition score evaluated by an expert, and the vertical axis is the average slope calculated by the calculation unit 34. The multiple black circles in FIG. 14 show the results of each dairy cow that underwent the experiment, and the dotted straight line shows an approximation line (regression line). As shown in FIG. 14, even if the range from the top point 91 of the contour line 60 to the inside of the contour line 60 by 20% to 30% of the reference dimension W is specified as the reference range 306a, the body condition score changes depending on the slope. The larger the negative value of the slope, the lower the body condition score, and the larger the positive value of the slope, the higher the body condition score. The approximation line was found to be y = 0.1868x - 0.5657, and the coefficient of determination ^R2 was 0.6210. From this, it can be seen that the beefiness of a cow can be evaluated with high accuracy even when the reference range 306a is specified as the range from the top point 91 of the contour line 60 to the inside of the contour line 60, which is 20% to 30% of the reference dimension W.

[Regarding reference range 306a]
In the above-described second embodiment, the reference range 306a is defined as a range from the uppermost point 91 of the contour line 60 to the inside of the contour line 60, the range being 20% to 30% of the reference dimension W. Here, a description will be given of how far the reference range 306a should be defined as being from the uppermost point 91 of the contour line 60.

[experiment]
In defining reference range 306a as the range away from uppermost point 91 of contour line 60 to the inside of contour line 60 by a first percentage or more and a second percentage or less of the reference dimension W, a plurality of reference ranges 306a were specified by changing the first percentage and the second percentage to various values. Then, as in Fig. 14, a correlation was found between the body condition score evaluated by a veterinarian or other expert and the slope of a line passing through two of the plurality of points 307a on contour line 60 in each of the plurality of reference ranges 306a, and a coefficient of determination ^R2 was calculated from the approximation line indicating the correlation.

Table 4 shows the coefficient of determination ^R2 obtained from an approximation line showing the correlation between the body condition score and the slope of the line for each of a plurality of reference ranges 306a with different distances from the top point 91 of the contour line ^60. The reference ranges 306a on the left (lower) side of Table 4 were obtained from an approximation line showing the correlation between the body condition score and the slope of a line passing through two points on the contour line 60 at both ends of the reference range 306a. The reference ranges 306a on the center and right (upper) side of the reference ranges 306a and the coefficient of determination ^R2 were obtained from an approximation line showing the correlation between the body condition score and the average slope of a plurality of lines passing through two adjacent points among the plurality of points 307 on the contour line 60 located at an interval of 0.05W within the reference range 306a.

As shown in Table 4, a large coefficient of determination ^R2 was obtained when the reference range 306a was set to 0.2 W to 0.3 W and 0.15 W to 0.35 W. From this, it can be seen that when the reference range 306a is set to a range away from the uppermost point 91 of the contour line 60 by a first percentage or more and a second percentage or less of the reference dimension W, a large coefficient of determination ^R2 can be obtained by setting the first percentage to a value of 15% or more and the second percentage to a value of 35% or less.

Therefore, when the determination unit 38 specifies the reference range 306a between the straight line 330a that is a first distance away from the top point 91 of the contour line 60 inside the contour line 60, which corresponds to the first ratio of the reference dimension W, and the straight line 332a that is a second distance away from the top point 91 of the contour line 60, which corresponds to the second ratio of the reference dimension W, it is preferable to use a value of 15% or more and a value of 35% or less as the second ratio. This allows the beefiness of the cow to be evaluated with high accuracy. In order to accurately evaluate the beefiness of the cow, it is preferable to use a value of 15% or more and 25% or less as the first ratio, more preferably a value of 17% or more and 23% or less, and even more preferably a value of 19% or more and 21% or less. It is preferable to use a value of 25% or more and 35% or less as the second ratio, more preferably a value of 27% or more and 33% or less, and even more preferably a value of 29% or more and 31% or less. Additionally, the width of the reference range 306a in the vertical direction (the cow's height direction) within the extracted image 51 is preferably 0.05W or more, more preferably 0.08W or more, and even more preferably 0.1W or more.

In the above second embodiment, as shown in FIG. 13, a case has been shown in which multiple straight lines 312a, 312b are determined from multiple points 307a within the reference range 306a on the left side of the contour 60, but straight lines may also be determined from multiple points within the reference range 306a on the right side of the contour 60. When the right side of the contour 60 is used, the relationship between the positive and negative slope of the line passing through multiple points on the contour 60 within the reference range 306a and the body condition score is opposite to that when the left side of the contour 60 is used. In other words, the more positive the slope of the line, the lower the body condition score, and the more negative the slope, the higher the body condition score.

The evaluation unit 36 may combine the first and second embodiments and evaluate the beefiness of the cow based on the average value of either or both of the body condition score obtained from the slope of one or more straight lines determined in the reference range 306 determined from the leftmost point 302 and the body condition score obtained from the slope of one or more straight lines determined in the reference range 306 determined from the rightmost point 303, and the body condition score obtained from one or more straight lines determined in the reference range 306a determined from the topmost point 91.

In the first and second embodiments, the determination unit 38 may specify the distance between the apexes 62a, 62b of the peaks 61a, 61b as the reference dimension indicating the width of the cow's body, or may specify the distance between the uppermost points of the peaks 61a, 61b.

The apex 62a of the mountain 61a and the apex 62b of the mountain 61b can be identified by the following method. For example, as shown in FIG. 5, an inclined straight line 55 inclined from the top to the bottom left in the extracted image 51 is moved from the top left end point 52 of the extracted image 51 toward the contour line 60, and the point where the inclined straight line 55 first touches the contour line 60 can be identified as the apex 62a of the mountain 61a on the left side. Similarly, an inclined straight line 56 inclined from the top to the bottom right in the extracted image 51 is moved from the top right end point 54 of the extracted image 51 toward the contour line 60, and the point where the inclined straight line 56 first touches the contour line 60 can be identified as the apex 62b of the mountain 61b on the right side. The inclined straight lines 55 and 56 are preferably inclined at 30° to 60° from the horizontal line in the extracted image 51, more preferably at 40° to 50°, and even more preferably at 45°.

The apex 62a of the mountain portion 61a and the apex 62b of the mountain portion 61b may be identified by other methods, not limited to the above-mentioned method. For example, as shown in FIG. 5, a first point of contact may be identified by approaching a slanted straight line 55a inclined from the top to the bottom left of the extracted image 51 from the upper left end point 52 toward the contour line 60, and a second point of contact may be identified by approaching a slanted straight line 55b inclined at a second predetermined angle from the upper left end point 52 toward the contour line 60, and a point located approximately in the center between the first point and the second point on the contour line 60 may be identified as the apex 62a of the mountain portion 61a. Similarly, a third point where an inclined straight line 56a inclined from the top to the bottom right in the extracted image 51 first touches the contour line 60 from the upper right end point 54 may be approached, and a fourth point where an inclined straight line 56b inclined at a fourth predetermined angle may be approached from the upper right end point 54 to the contour line 60 to identify the fourth point where the first touches the contour line 60. A point on the contour line 60 located approximately in the middle between the third point and the fourth point may be identified as the apex 62b of the mountain portion 61b. The first and third predetermined angles are preferably 15° to 30° from the horizontal line in the extracted image 51, and the second and fourth predetermined angles are preferably 60° to 75° from the horizontal line in the extracted image 51.

In addition, for example, the extracted image 51 may be converted into an X-Y coordinate system, and the vertices 62a, 62b of the two peaks 61a, 61b may be identified from the X and Y coordinates of each pixel of the extracted image 51. For example, the point on the contour 60 that is the shortest distance from the upper left corner 52 of the extracted image 51 may be set as the vertex 62a of the peak 61a, and the point on the contour 60 that is the shortest distance from the upper right corner 54 of the extracted image 51 may be set as the vertex 62b of the peak 61b.

Alternatively, the apex of either the peak 61a or the peak 61b may be specified by any of the above methods, and the point where the horizontal line in the extracted image 51 passing through the specified apex intersects with the contour line 60 of the other peak may be specified as the apex of the other peak. Note that the apex 62a of the peak 61a and the apex 62b of the peak 61b may be specified by a method other than the above methods. Note that the range of the peaks 61a and 61b can be specified by various methods. For example, the range of the peak 61a can be the range of the contour line 60 contained in a circle of a predetermined radius centered on the apex 62a of the peak 61a. The predetermined radius can be, for example, a value obtained by multiplying the distance between the apex 62a of the peak 61a and the uppermost point 91 located at the uppermost side of the contour line 60 by a predetermined value (for example, 0.5), but may be a value obtained by other methods. Also, the range of the peak 61a may be a range of a predetermined length on both sides from the apex 62a of the peak 61a. The predetermined length may be, for example, the length of the above-mentioned predetermined radius, or may be the length of the entire contour line 60 multiplied by a predetermined value. The same applies to the peak portion 61b. The range of the peak portions 61a and 61b may also be determined by machine learning or the like.

In the above first and second embodiments, the image acquisition unit 30 has been described as acquiring an image of a cow in a standing position taken from approximately directly behind, but it may also be possible to acquire an image of a cow in a standing position taken from approximately the front. FIG. 15(a) is an image 50 of a cow in a standing position taken from approximately directly behind, and FIG. 15(b) is an image 50a of the cow in FIG. 15(a) taken from approximately the front. As shown in FIG. 15(a), the image is not limited to an image of a cow in a standing position taken from approximately directly behind. Even if an image of a cow in a standing position taken from approximately the front, as shown in FIG. 15(b), it is possible to image the cow so that the contours around the left and right hip angles of the cow are visible.

16(a) and 16(b) are extracted images 51 in which the contour line 60 of the cow is extracted from the captured images 50 and 50a of FIG. 15(a) and FIG. 15(b). As shown in FIG. 16(a), even if the image of the cow in a standing position is taken from almost directly behind, as shown in FIG. 16(b), even if the image of the cow in a standing position is taken from almost directly in front, the ridges 61a and 61b corresponding to the waist angle can be present on both the left and right sides of the upper part of the contour line 60. Therefore, even if the image 50a taken of the cow in a standing position from almost directly in front is used, the beefiness of the cow can be evaluated in the same manner as when the image 50 taken from almost directly behind is used. In addition, even if the two-dimensional captured image 50a taken of the cow in a standing position from almost directly in front is used, as in the case of using the image 50 taken from almost directly behind, no special equipment is required to capture the image of the cow from above, and it can be applied to various breeding sites, and a portable information device such as a smartphone, which is highly portable and versatile, can be used as the evaluation device.

In the above embodiments, the evaluation device for evaluating the beefiness of the cattle is a portable information device with an image capturing unit 18 that is carried by the worker, but other cases are also possible. For example, it may be a stationary information device such as a server that can communicate with other devices connected to a network. In this case, the worker transmits the image of the cattle captured using a portable information device such as a smartphone to a stationary information device such as a server via the network. The stationary information device performs the same process as the evaluation device 100 to evaluate the beefiness of the cattle. The stationary information device then transmits the evaluation result of the beefiness of the cattle to a portable information device carried by the worker via the network. This reduces the load on the smartphone. In addition, a fixed or movable camera may be installed in the cattle barn, and the portable information device such as a smartphone and the stationary information device such as a server may acquire an image of the cattle standing upright, captured by the camera installed in the cattle barn, from approximately directly behind or approximately in front of the cattle.

In the above first and second embodiments, the body condition score is evaluated as the amount of meat in the cow, but other indices for evaluating the amount of meat in the cow may be evaluated. In addition, although each experiment was performed on dairy cows, it is believed that beef cows, which have a similar skeletal structure, would produce similar results. Therefore, each of the above embodiments can be applied not only to dairy cows but also to beef cows.

Although the embodiment of the present invention has been described in detail above, the present invention is not limited to such a specific embodiment, and various modifications and variations are possible within the scope of the gist of the present invention as described in the claims.

12 Display unit 14 Operation unit 16 Communication unit 18 Imaging unit 20 Memory unit 22 Control unit 30 Image acquisition unit 32 Extraction unit 34 Calculation unit 36 Evaluation unit 38 Determination unit 50, 50a Captured image 51 Extracted image 52 Upper left end point 54 Upper right end point 55, 55a, 55b, 56, 56a, 56b Inclined straight line 60 Contour line 61a, 61b Mountain portion 62a, 62b Vertex 63a, 63b First curved portion 64a, 64b Second curved portion 91 Uppermost point 100 Evaluation device 302 Leftmost point 303 Rightmost point 304, 305 Vertical line 306, 306a Reference range 307, 307a Point 310, 310a, 310b, 310c, 310d, 310e straight line 312a, 312b straight line 330, 330a, 332, 332a straight line

Claims (11)

an image acquisition unit for acquiring a two-dimensional image of a cow in a standing position from approximately directly behind or approximately directly in front of the cow;
an extraction unit that extracts a contour line of the cow from the captured image;
a determination unit that specifies a reference dimension related to the size in the width direction of the body of the cow from the contour line, specifies a reference range between two straight lines that are separated from the leftmost point on the left side, the rightmost point on the right side, or the topmost point on the top side of the contour line by a first distance corresponding to a first ratio of the reference dimension and a second distance corresponding to a second ratio of the reference dimension inside the contour line and are approximately perpendicular to the direction of separation, and determines one or more straight lines from a plurality of points on a continuous portion of the contour line that is located within the reference range;
a calculation unit that calculates a slope of the one or more straight lines determined by the determination unit or a value obtained from the slope;
and an evaluation unit that evaluates the fleshiness of the cow from the slope calculated by the calculation unit or a value obtained from the slope. The cattle evaluation device according to claim 1, wherein the determination unit determines the distance between a straight line that passes through the leftmost point of the contour line in the extracted image from which the contour line is extracted and extends vertically, and a straight line that passes through the rightmost point of the contour line in the extracted image and extends vertically, as the reference dimension in the width direction of the body of the cattle. The cattle evaluation device according to claim 2, wherein the determination unit identifies the reference range located between the two straight lines that are separated from the leftmost point or the rightmost point of the contour line by the first distance and the second distance inside the contour line, respectively, and uses a value of 16% or more as the first ratio and a value of 40% or less as the second ratio. The cattle evaluation device according to claim 3, wherein the determination unit uses a value between 16% and 24% as the first ratio and a value between 28% and 40% as the second ratio. The cattle evaluation device according to claim 2, wherein the determination unit identifies the reference range located between the two straight lines that are separated from the top point of the contour line by the first distance and the second distance inside the contour line, and uses a value of 15% or more as the first ratio and a value of 35% or less as the second ratio. The cattle evaluation device according to claim 5, wherein the determination unit uses a value of 15% or more and 25% or less as the first ratio, and a value of 28% or more and 35% or less as the second ratio. The cattle evaluation device according to any one of claims 1 to 6, wherein the determination unit determines, as the one or more straight lines, a single straight line passing through points on the contour line at both ends of the reference range. the determining unit determines, as the one or more straight lines, a plurality of straight lines from the plurality of points on a continuous portion of the contour line that is located within the reference range;
The cow evaluation device according to claim 1 , wherein the calculation unit calculates a maximum slope, a minimum slope, or an average slope of the plurality of straight lines. The cattle evaluation device according to any one of claims 1 to 8, wherein the evaluation unit evaluates the body condition score as an evaluation of the beefiness of the cattle. The cattle evaluation device according to any one of claims 1 to 9, comprising an imaging unit that images the cattle. A contour line of a cow is extracted from a two-dimensional captured image of the cow in a standing position taken from approximately directly behind or approximately directly in front of the cow;
Identifying a reference dimension related to the size of the cow's body in the width direction from the extracted contour line, identifying a reference range sandwiched between two straight lines that are separated from the leftmost point on the leftmost side, the rightmost point on the rightmost side, or the topmost point on the topmost side of the contour line by a first distance corresponding to a first ratio of the reference dimension and a second distance corresponding to a second ratio of the reference dimension inside the contour line and are approximately perpendicular to the direction of separation, and determining one or more straight lines from a plurality of points on a continuous portion of the contour line that is located within the reference range,
Calculating a slope of the one or more determined straight lines or a value derived from the slope;
The meatiness of the cattle is evaluated from the calculated slope or a value obtained from the slope.
A method for evaluating cattle, characterized in that processing is carried out by a computer.

Images