Animal behavior analysis method and apparatus based on animal postures, and device and medium

Abstract

The present application relates to the technical field of animal behavior analysis. Disclosed are an animal behavior analysis method and apparatus based on animal postures, and a device and a medium. The method comprises: acquiring animal images at a plurality of angles of view; performing posture analysis on the animal images, so as to obtain animal posture feature images, and performing action recognition on the basis of the animal posture feature images, so as to obtain action types; performing normalized coordinate system conversion on the basis of three-dimensional coordinates of the animal posture feature images, so as to obtain target position data; performing analysis on the basis of the action types, time information respectively corresponding to the animal images, and the target position data, so as to determine whether the behavior of an animal in the animal images is abnormal. Accurate analysis can be performed by means of action types of extracted animal posture feature images, time information, and target position data, so as to determine whether the behavior of an animal in animal images is abnormal, thereby improving the effect of analyzing the behavior of the animal.

Description

Animal behavior analysis method, device, equipment and medium based on animal posture Technical Field

[0001] The present invention relates to the technical field of animal behavior analysis based on animal posture, and in particular to an animal behavior analysis method, device, equipment and medium based on animal posture. Background Art

[0002] For a long time, the emotional and health status of laboratory animals has typically been assessed using low-dimensional physical parameters such as movement distance, time spent in a central area, and interaction distance in behavioral tests. However, it is unclear what postures the animals maintain in different positions, what movements they make, and what behavioral and emotional states they represent. For example, stopping in one place could mean sleeping or grooming. Therefore, finding a more accurate way to analyze animal behavior has become an urgent problem.

[0003] Summary of the Invention

[0004] Based on this, it is necessary to address the technical problem that the existing technology has poor results in analyzing animal behavior, and propose an animal behavior analysis method, device, equipment and medium based on animal posture.

[0005] In a first aspect, a method for analyzing animal behavior based on animal posture is provided, the method comprising:

[0006] Acquire images of various animals from multiple perspectives, wherein the animal images are obtained by photographing the target animal in a target environment using a target camera device;

[0007] Performing posture analysis on each of the animal images to obtain a posture feature image of each animal, and performing action recognition based on the posture feature image of the animal and a trained action behavior recognition model to obtain an action type corresponding to each of the animal images;

[0008] Performing normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture feature image to obtain target position data of each of the animal posture feature images in the target coordinate system;

[0009] An analysis is performed based on the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images on the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal.

[0010] In a second aspect, an animal behavior analysis device based on animal posture is provided, the device comprising:

[0011] An acquisition module, configured to acquire images of various animals from multiple viewing angles, wherein the images of the animals are obtained by photographing the target animals in a target environment using a target camera device;

[0012] a posture analysis module for performing posture analysis on each of the animal images to obtain a posture feature image of each animal, and performing action recognition based on the posture feature image of the animal and a trained action behavior recognition model to obtain the action type corresponding to each of the animal images;

[0013] A coordinate conversion module, configured to perform normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture characteristic image to obtain target position data of each of the animal posture characteristic images in the target coordinate system;

[0014] The abnormality module is used to analyze the action type corresponding to each animal image, the time information corresponding to each animal image, and the target position data of each animal posture feature image in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal.

[0015] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned animal behavior analysis method based on animal posture when executing the computer program.

[0016] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned animal behavior analysis method based on animal posture are implemented.

[0017] The present invention proposes an animal behavior analysis method based on animal posture, which obtains animal images from multiple perspectives, wherein the animal images are obtained by photographing a target animal in a target environment using a target camera device. Then, posture analysis is performed on each animal image to obtain each animal posture feature image. Action recognition is performed based on the animal posture feature image and a trained action behavior recognition model to obtain the action type corresponding to each animal image. Then, normalized coordinate system transformation is performed based on the three-dimensional coordinates of the animal posture feature image to obtain target position data of each animal posture feature image in the target coordinate system. Finally, analysis is performed based on the action type corresponding to each animal image, the time information corresponding to each animal image, and the target position data of each animal posture feature image in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal. The action type, time information, and target position data corresponding to the animal posture feature image can be obtained through the extracted animal posture feature image. The action type, time information, and target position data are accurately analyzed to determine whether the behavior of the animal in the animal image is abnormal, thereby improving the effect of animal behavior analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] in:

[0020] FIG1 is a diagram illustrating an application environment of an animal behavior analysis method based on animal posture in one embodiment;

[0021] FIG2 is a flow chart of an animal behavior analysis method based on animal posture according to one embodiment;

[0022] FIG3 is a skeleton image of a rat in an animal state analysis method according to an embodiment;

[0023] FIG4 is a block diagram of an animal behavior analysis device based on animal posture according to one embodiment;

[0024] FIG5 is a block diagram of a computer device according to an embodiment;

[0025] FIG6 is a block diagram of a computer device according to another embodiment;

[0026] FIG7 is a schematic diagram of a target heat map of an animal behavior analysis method based on animal posture in one embodiment. DETAILED DESCRIPTION

[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0028] The animal behavior analysis method based on animal posture provided by the embodiment of the present invention can be applied in an application environment as shown in FIG. 1 , wherein the client 110 communicates with the server 120 via a network. The server 120 can receive various animal images from multiple perspectives through the client 110, wherein the animal images are obtained by shooting the target animal with a target camera device in a target environment. The server 120 then performs posture analysis on each of the animal images to obtain each animal posture feature image, and performs action recognition based on the animal posture feature image and a trained action behavior recognition model to obtain the action type corresponding to each of the animal images. The server 120 then performs normalized coordinate system conversion based on the three-dimensional coordinates of the animal posture feature image to obtain the target position data of each of the animal posture feature images in the target coordinate system. Finally, the server 120 analyzes the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal. The server 120 can obtain the action type, time information, and target position data corresponding to the animal posture feature image through the extracted animal posture feature images, and accurately analyze the action type, time information, and target position data, thereby determining whether the behavior of the animal in the animal image is abnormal, thereby improving the effect of analyzing the animal's behavior. The client 110 may be, but is not limited to, various personal computers, laptop computers, smartphones, tablet computers, and portable wearable devices. The server 120 may be implemented as an independent server or a server cluster consisting of multiple servers. The present invention will be described in detail below through specific embodiments.

[0029] Please refer to FIG2 , which is a flow chart of an animal behavior analysis method based on animal posture according to an embodiment of the present invention, including the following steps:

[0030] Step S101: acquiring images of animals from multiple perspectives, wherein the images of animals are obtained by photographing a target animal in a target environment using a target camera device;

[0031] In a preferred embodiment, the step of obtaining images of each animal from multiple perspectives includes: obtaining each captured image, wherein the captured images are obtained by capturing different target camera devices in a target environment; and among each of the captured images, selecting the captured image in which the target object in the captured image matches the target animal as the animal image.

[0032] In a preferred embodiment, multiple cameras are used to shoot animals from multiple angles in a target environment, and matching captured images are screened out as animal images for specific animal postures, so as to quickly locate the time and location of the animal's action.

[0033] Among them, the target animals can be mice, cats, dogs and other animals, and the target environment can be an open field, a pet hospital, a home environment and the like.

[0034] Step S102: performing posture analysis on each of the animal images to obtain a posture feature image of each animal, and performing action recognition based on the posture feature image of the animal and a trained action behavior recognition model to obtain an action type corresponding to each of the animal images;

[0035] For example, for each animal image in the animal image set, an animal posture feature image is generated, and each animal posture feature image is used as the animal posture feature image set. As shown in FIG3 , the animal posture feature image includes a white rat skeleton image when the animal is a white rat.

[0036] As an example, for each animal image, the two-dimensional spatial coordinates of 16 body key points of each frame of the animal image are extracted, and the three-dimensional skeleton of the animal is constructed according to the camera calibration file and the above two-dimensional spatial coordinates to obtain an animal posture feature image, wherein the body key points include at least the four claws of the animal in the animal image; then the animal posture feature images corresponding to each animal image are aligned, and the body orientation is unified; the animal skeletons in the animal posture feature images are scaled to the same scale size; the active horizontal plane of the animal skeleton in the animal posture feature image is corrected so that the animal skeleton in the proposed animal posture feature image moves on the same horizontal base plane, and each corrected animal posture feature image is used as an animal posture feature image set.

[0037] For example, the action recognition model can be a classification model trained using a convolutional neural network, which can utilize an unsupervised algorithm. Action recognition is performed by inputting animal posture feature images into the trained action recognition model. The model then outputs the action type corresponding to each animal image. Action types may include sniffing, grooming, eating, and nesting.

[0038] Step S103: performing normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture characteristic image to obtain target position data of each of the animal posture characteristic images in the target coordinate system;

[0039] In this embodiment, since the animal posture characteristic images can be obtained by shooting with different target camera devices, the camera coordinate systems of each target camera device are different due to different angles or positions, so the animal posture characteristic images are converted into a normalized coordinate system, thereby obtaining the target position data of each of the animal posture characteristic images in the normalized coordinate system. The normalized coordinate system can be pre-set manually. In one implementation method, the camera coordinate system of a target camera device when shooting an image is used as the normalized coordinate system, that is, the optical center of the camera when shooting an image is used as the origin, the X-axis is the horizontal direction, the Y-axis is the vertical direction, and the Z-axis points to the direction observed by the camera when shooting an image. After selection, the target coordinate system will no longer change, that is, it is unchanged and unique.

[0040] Furthermore, in a preferred embodiment, the step of performing normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture feature image to obtain target position data of each of the animal posture feature images in the target coordinate system includes:

[0041] Step S201: for each of the animal posture feature images, convert the three-dimensional coordinates of the animal posture feature image into a normalized coordinate system to obtain target position data of the animal posture feature image in the normalized coordinate system.

[0042] For example, the three-dimensional coordinates of the animal posture feature image are converted into a normalized coordinate system to obtain position data of the animal posture feature image in the normalized coordinate system, and the position data is used as the target position data.

[0043] Furthermore, in a preferred embodiment, for each of the animal posture feature images, the step of converting the three-dimensional coordinates of the animal posture feature image into a normalized coordinate system to obtain the target position data of the animal posture feature image in the normalized coordinate system includes:

[0044] Step S301: extracting the three-dimensional coordinates of the key points on the back of the animal in the animal posture feature image from the three-dimensional coordinates of each of the animal posture feature images as target posture position data;

[0045] Step S302: converting the target posture position data into a normalized coordinate system to obtain the target position data of the animal posture feature image in the normalized coordinate system.

[0046] In this embodiment, in order to improve the accuracy of the target position data of the animal posture feature image in the normalized coordinate system, each key point on the animal posture feature image corresponds to a bone position data, and the bone position data is a three-dimensional coordinate; the bone position data of the key point corresponding to the back of the animal in the animal posture feature image is used as the target posture position data.

[0047] Step S104: Analyze the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal.

[0048] The time information refers to the time when the animal image is captured, and the time may be a timestamp.

[0049] For example, in an open field environment, mice in the open field have some specific habits, such as foraging at a certain place at a fixed time and place. The experimenters collect data on the daily behavioral habits of mice in the open field; when the animal in the animal image is a mouse in the open field, the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system can be compared with the daily behavioral habits of the mice in the open field to determine whether the behavior of the animal in the animal image is abnormal.

[0050] As an example, when the behavior of an animal in an animal image is abnormal, an alarm signal is generated and sent to a user terminal to alert the user and enable the user to take corresponding measures in a timely manner.

[0051] Furthermore, in a preferred embodiment, the step of analyzing the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal includes:

[0052] Step S401: generating a heat map based on the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system to obtain a target heat map;

[0053] Step S402: performing comparative analysis based on the normalized historical thermogram and the target thermogram to determine whether the behavior of the animal in the animal image is abnormal.

[0054] The target heat map is shown in FIG7 .

[0055] The animal behavior analysis method based on animal posture proposed in this embodiment obtains animal images from multiple perspectives, wherein the animal images are obtained by photographing a target animal using a target camera device in a target environment. Then, posture analysis is performed on each of the animal images to obtain each animal posture feature image. Action recognition is performed based on the animal posture feature images and a trained action behavior recognition model to obtain the action type corresponding to each of the animal images. Then, normalized coordinate system transformation is performed based on the three-dimensional coordinates of the animal posture feature images to obtain target position data of each of the animal posture feature images in the target coordinate system. Finally, analysis is performed based on the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal. The action type, time information, and target position data corresponding to the animal posture feature images can be obtained from the extracted animal posture feature images. The action type, time information, and target position data can be accurately analyzed to determine whether the behavior of the animal in the animal image is abnormal, thereby improving the effectiveness of animal behavior analysis.

[0056] Referring to FIG. 4 , in one embodiment, an animal behavior analysis device based on animal posture is provided, the device comprising: an acquisition module 10 for acquiring images of animals from multiple perspectives, wherein the animal images are obtained by photographing a target animal in a target environment using a target camera device;

[0057] The posture analysis module 20 is used to perform posture analysis on each of the animal images to obtain each animal posture feature image, and perform action recognition based on the animal posture feature image and a trained action behavior recognition model to obtain the action type corresponding to each of the animal images.

[0058] A coordinate conversion module 30 is used to perform normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture characteristic image to obtain target position data of each of the animal posture characteristic images in the target coordinate system;

[0059] The abnormality module 40 is used to analyze the action type corresponding to each animal image, the time information corresponding to each animal image, and the target position data of each animal posture feature image in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal.

[0060] Furthermore, in one embodiment, the coordinate conversion module 30 is used to: for each of the animal posture feature images, convert the three-dimensional coordinates of the animal posture feature image into a normalized coordinate system to obtain the target position data of the animal posture feature image in the normalized coordinate system.

[0061] Furthermore, in one embodiment, the coordinate conversion module 30 is configured to: extract, from the three-dimensional coordinates of each of the animal posture feature images, the three-dimensional coordinates of the key points on the back of the animal in the animal posture feature image as the target posture position data;

[0062] The target posture position data is converted into a normalized coordinate system to obtain the target position data of the animal posture feature image in the normalized coordinate system.

[0063] Furthermore, in one embodiment, the abnormality module 40 is also used to: generate a heat map based on the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system to obtain a target heat map; and perform comparative analysis based on the normalized historical heat map and the target heat map to determine whether the behavior of the animal in the animal image is abnormal.

[0064] Furthermore, in one embodiment, the acquisition module 10 is further configured to: acquire each captured image, wherein the captured image is captured by different target camera devices in a target environment;

[0065] Among the captured images, the captured images in which the target object matches the target animal are selected as animal images.

[0066] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in FIG5 . The computer device includes a processor, a memory, a network interface, and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it implements the functions or steps of the server side of an animal behavior analysis method based on animal posture.

[0067] In one embodiment, a computer device is provided, which may be a client, and its internal structure diagram may be as shown in FIG6 . The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements the functions or steps on the client side of an animal behavior analysis method based on animal posture.

[0068] In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented:

[0069] Acquire images of various animals from multiple perspectives, wherein the animal images are obtained by photographing the target animal in a target environment using a target camera device;

[0070] Performing posture analysis on each of the animal images to obtain a posture feature image of each animal, and performing action recognition based on the posture feature image of the animal and a trained action behavior recognition model to obtain an action type corresponding to each of the animal images;

[0071] Performing normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture feature image to obtain target position data of each of the animal posture feature images in the target coordinate system;

[0072] An analysis is performed based on the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images on the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal.

[0073] The animal behavior analysis method based on animal posture proposed in this embodiment obtains animal images from multiple perspectives, wherein the animal images are obtained by photographing a target animal using a target camera device in a target environment. Then, posture analysis is performed on each of the animal images to obtain each animal posture feature image. Action recognition is performed based on the animal posture feature images and a trained action behavior recognition model to obtain the action type corresponding to each of the animal images. Then, normalized coordinate system transformation is performed based on the three-dimensional coordinates of the animal posture feature images to obtain target position data of each of the animal posture feature images in the target coordinate system. Finally, analysis is performed based on the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal. The action type, time information, and target position data corresponding to the animal posture feature images can be obtained from the extracted animal posture feature images. The action type, time information, and target position data can be accurately analyzed to determine whether the behavior of the animal in the animal image is abnormal, thereby improving the effectiveness of animal behavior analysis.

[0074] In one embodiment, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0075] Acquire images of various animals from multiple perspectives, wherein the animal images are obtained by photographing the target animal in a target environment using a target camera device;

[0076] Performing posture analysis on each of the animal images to obtain a posture feature image of each animal, and performing action recognition based on the posture feature image of the animal and a trained action behavior recognition model to obtain an action type corresponding to each of the animal images;

[0077] Performing normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture feature image to obtain target position data of each of the animal posture feature images in the target coordinate system;

[0078] An analysis is performed based on the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images on the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal.

[0079] The animal behavior analysis method based on animal posture proposed in this embodiment obtains animal images from multiple perspectives, wherein the animal images are obtained by photographing a target animal using a target camera device in a target environment. Then, posture analysis is performed on each of the animal images to obtain each animal posture feature image. Action recognition is performed based on the animal posture feature images and a trained action behavior recognition model to obtain the action type corresponding to each of the animal images. Then, normalized coordinate system transformation is performed based on the three-dimensional coordinates of the animal posture feature images to obtain target position data of each of the animal posture feature images in the target coordinate system. Finally, analysis is performed based on the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each of the animal posture feature images in the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal. The action type, time information, and target position data corresponding to the animal posture feature images can be obtained from the extracted animal posture feature images. The action type, time information, and target position data can be accurately analyzed to determine whether the behavior of the animal in the animal image is abnormal, thereby improving the effectiveness of animal behavior analysis.

[0080] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can be found in the relevant descriptions of the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0081] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0082] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0083] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. An animal behavior analysis method based on animal postures, characterized in that, the animal behavior analysis method based on animal postures includes: acquiring each animal image from multiple perspectives, where the animal image is obtained by a target imaging device photographing a target animal in a target environment; performing posture analysis on each of the animal images to obtain each animal posture feature image, and performing action recognition according to the animal posture feature image and a trained action behavior recognition model to obtain the action type corresponding to each of the animal images; performing normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture feature image to obtain the target position data of each animal posture feature image on the target coordinate system; analyzing according to the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each animal posture feature image on the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal.

2. The animal behavior analysis method based on animal postures according to claim 1, characterized in that, the step of performing normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture feature image to obtain the target position data of each animal posture feature image on the target coordinate system includes: for each animal posture feature image, converting the three-dimensional coordinates of the animal posture feature image into a normalized coordinate system to obtain the target position data of the animal posture feature image on the normalized coordinate system.

3. The animal behavior analysis method based on animal postures according to claim 2, characterized in that, the step of for each animal posture feature image, converting the three-dimensional coordinates of the animal posture feature image into a normalized coordinate system to obtain the target position data of the animal posture feature image on the normalized coordinate system includes: in the three-dimensional coordinates of each animal posture feature image, extracting the three-dimensional coordinates of the key points corresponding to the back of the animal in the animal posture feature image as the target posture position data; converting the target posture position data into a normalized coordinate system to obtain the target position data of the animal posture feature image on the normalized coordinate system.

4. The animal behavior analysis method based on animal postures according to claim 1, characterized in that, the step of analyzing according to the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each animal posture feature image on the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal includes: generating a target heat map according to the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each animal posture feature image on the target coordinate system; performing comparative analysis based on the normalized historical heat map and the target heat map to determine whether the behavior of the animal in the animal image is abnormal.

5. The method for analyzing animal behavior based on animal posture according to claim 1, wherein, the step of obtaining each animal image from multiple perspectives includes: obtaining each captured image, wherein the captured image is captured by different target imaging devices in a target environment; in each of the captured images, selecting the captured image in which the target object in the captured image matches the target animal as the animal image.

6. An apparatus for analyzing animal behavior based on animal posture, wherein, the apparatus for analyzing animal behavior based on animal posture includes: an acquisition module, configured to obtain each animal image from multiple perspectives, wherein the animal image is captured by a target imaging device for a target animal in a target environment; a posture analysis module, configured to perform posture analysis on each of the animal images to obtain each animal posture feature image, and perform action recognition according to the animal posture feature image and a trained action behavior recognition model to obtain the action type corresponding to each of the animal images.

7. The apparatus for analyzing animal behavior based on animal posture according to claim 6, wherein, the apparatus for analyzing animal behavior based on animal posture includes: a coordinate conversion module, configured to perform normalized coordinate system conversion according to the three-dimensional coordinates of the animal posture feature image to obtain the target position data of each animal posture feature image on the target coordinate system; an anomaly module, configured to analyze according to the action type corresponding to each of the animal images, the time information corresponding to each of the animal images, and the target position data of each animal posture feature image on the target coordinate system to determine whether the behavior of the animal in the animal image is abnormal.

8. A computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein, when the processor executes the computer program, the steps of the method for analyzing animal behavior based on animal posture according to any one of claims 1 to 5 are implemented.

9. A computer-readable storage medium storing a computer program, wherein, when the computer program is executed by a processor, the steps of the method for analyzing animal behavior based on animal posture according to any one of claims 1 to 5 are implemented.

Images