Northeast tiger individual left and right image association method based on spatio-temporal information

Abstract

A method for associating left and right images of an individual northeast tiger based on spatio-temporal information belongs to the field of animal identification, and comprises the following steps: collecting and marking data of the individual northeast tiger shot by a wild protection camera; predicting key point information and body side orientation of the northeast tiger individuals based on a key point detection algorithm; estimating individual information of the northeast tiger based on an individual identification algorithm; and associating northeast tiger individuals in the left and right body side image libraries in combination with spatio-temporal information. According to the method, the northeastern tiger individuals after left and right side image association have bilateral information, and the correct northeastern tiger individuals can be accurately matched even if only one side is shot in the field, so that the detection precision is improved; by associating the left and right images of the northeast tiger individuals and fully utilizing the spatio-temporal information of the northeast tiger, a solid data basis can be provided for population dynamic monitoring, the population number can be estimated more accurately, and the home domain range can be measured and calculated; the method does not need to depend on artificial expert experience, detection time is shortened, multi-person cooperative detection operation can be achieved, and operation is convenient and fast.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of animal identification, and particularly relates to a northeast tiger individual left-right side image association method based on space-time information. BACKGROUND

[0002] The northeast tiger is one of the most endangered large cats in the world. Accurate identification and identity confirmation of each northeast tiger in nature is not only a cutting-edge scientific research practice, but also a core work related to species survival, ecological balance and protection effectiveness, which has crucial and far-reaching significance.

[0003] At present, the northeast tiger individual identification method based on computer vision usually utilizes the unique identification characteristics of the body side patterns of different northeast tiger individuals, extracts the features of the northeast tiger body side images based on a neural network, and then predicts individual information. This method regards the two side images of the same northeast tiger individual as two different individuals when identifying, and the obtained result cannot be directly used to analyze the population number and territory range of the northeast tiger. In order to solve this problem, two wildlife protection cameras can be deployed at each monitoring point in the wild, so that when the northeast tiger individual appears, the data of the body sides can be simultaneously captured. However, this method consumes huge equipment cost and deployment cost. Another method is to manually analyze the video, and according to the time, scene and appearance characteristics of the northeast tiger in the video, the left and right side images of the same individual are associated. However, this method depends on expert experience and is not efficient. Therefore, it is a technical problem to be solved by those skilled in the art to provide a method for automatically associating the left and right side images of the same northeast tiger individual in the deployment scene of a single wildlife protection camera.

[0004] The existing northeast tiger individual identification method has the following problems:

[0005] (1) The left side image and the right side image of the same northeast tiger individual are regarded as different individuals, resulting in low detection accuracy;

[0006] (2) Manual association depends on expert experience, consumes time and cannot realize multi-person cooperation;

[0007] (3) The territory characteristics and space-time information of the northeast tiger when moving are not fully utilized. SUMMARY

[0008] In order to solve the problems of the existing northeast tiger individual identification method, such as regarding the left side image and the right side image of the same northeast tiger individual as different individuals, manual association depending on expert experience, consuming time and being unable to realize multi-person cooperation, and not fully utilizing space-time information, the application provides a northeast tiger individual left-right side image association method based on space-time information.

[0009] The technical solution adopted by the application to solve the technical problems is as follows:

[0010] The application provides a left-right side image association method for Siberian tiger individuals based on space-time information.

[0011] Step S1: collect and label data of Siberian tiger individuals photographed by a wildlife camera;

[0012] Step S2: predict key point information and body side orientation of the Siberian tiger individual based on a key point detection algorithm;

[0013] Step S3: estimate individual information of the Siberian tiger based on an individual recognition algorithm;

[0014] Step S4: associate left and right body side images of the Siberian tiger individual in a left and right body side image database based on space-time information.

[0015] Further, the specific implementation process of step S1 is as follows:

[0016] S1.1: collect images of Siberian tiger individuals photographed by a wildlife camera;

[0017] S1.2: label the position and key point information of the Siberian tiger individual in the Siberian tiger individual image, and set the individual number;

[0018] S1.3: count the time information of the Siberian tiger individual image photographed and the corresponding wildlife camera number;

[0019] S1.4: count the latitude and longitude information and attitude information of the installation position of the wildlife camera.

[0020] Further, the specific implementation process of step S2 is as follows:

[0021] S2.1: based on the labeled position and key point information of the Siberian tiger individual, construct a Siberian tiger key point detection data set;

[0022] S2.2: based on a key point detector, use the Siberian tiger key point detection data set to predict the body side image of the Siberian tiger individual, and obtain the position and key point information of the Siberian tiger individual;

[0023] S2.3: construct a Siberian tiger individual left side body image database and a Siberian tiger individual right side body image database according to the predicted body side image of the Siberian tiger individual.

[0024] Further, the labeled position information of the Siberian tiger individual is the information of the frame containing the Siberian tiger individual region.

[0025] Further, the key points of the labeled Siberian tiger individual have six in common: the key point 1 is the joint of the head and the body of the Siberian tiger individual; the key point 2 is the joint of the buttocks and the tail of the Siberian tiger individual; and the key points 3 to 6 are the joints of the left forelimb, the right forelimb, the left hind limb and the right hind limb of the Siberian tiger individual and the body, respectively.

[0026] Further, the body side image of the Siberian tiger individual is the left side body image of the Siberian tiger individual if the key point 1 is located on the left side of the key point 2, and the body side image of the Siberian tiger individual is the right side body image of the Siberian tiger individual if the key point 1 is located on the right side of the key point 2.

[0027] Further, the specific implementation process of step S3 is as follows:

[0028] S3.1: based on the position and key point information of the labeled Siberian tiger individual, a torso region image is cut out from the left side body image of the Siberian tiger individual, and the torso region image is associated with the individual number labeled;

[0029] S3.2: the right side body image in the Siberian tiger individual right side body image library is combined with the left side body image in the Siberian tiger individual left side body image library after horizontal flip, and the images in the combined image library are used as the Siberian tiger individual recognition training data set;

[0030] S3.3: an individual recognition network model is constructed, and the features of the Siberian tiger body side torso region image are extracted on the Siberian tiger individual recognition training data set, and the Siberian tiger individual information is predicted through the classification module.

[0031] Further, the specific implementation process of step S4 is as follows:

[0032] S4.1: the photographed Siberian tiger individual image is associated with the labeled individual number information and the time information of the photographed Siberian tiger individual, and a Siberian tiger space-time information library is constructed.

[0033] S4.2: the latitude and longitude information of the wildlife camera is obtained based on the positioning module inside the wildlife camera, and the relative distance matrix of the wildlife camera is calculated according to the latitude and longitude information , , where n represents the number of wildlife cameras, , represents the distance between the wildlife camera numbered and the wildlife camera numbered .

[0034] S4.3: based on the Siberian tiger space-time information library, an individual space-time matrix of the Siberian tiger individual left side body image is constructed Individual spatiotemporal matrix of the right side body image of an individual Siberian tiger , This represents the number of Siberian tiger individuals in the database containing images of the left side of their bodies. This represents the number of Siberian tiger individuals in the database containing images of the right side of their bodies. The number of weeks the wildlife camera worked. This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. Was the Siberian tiger individual numbered in week k? The value is 0 when the wildlife protection camera captures the image and 1 when the image is captured. This indicates that the image of the right side of an individual Siberian tiger, numbered [number missing], is located in the image database. The first Siberian tiger individual Is Zhou numbered? The value is 0 when the wildlife protection camera captures the image and 1 when the image is captured.

[0035] S4.4: Constructing the cost matrix based on the individual spatiotemporal matrix of the left side image of an individual Siberian tiger, the individual spatiotemporal matrix of the right side image of an individual Siberian tiger, and the relative distance matrix of the wildlife camera. Construct an extended cost matrix;

[0036] S4.5: Solve the extended cost matrix using the Hungarian matching algorithm.

[0037] Furthermore, the specific implementation process of step S4.4 is as follows:

[0038] By iterating through the spatiotemporal matrices of images of the left and right sides of an individual Siberian tiger over time, and at weekly intervals, the distance cost sets of all Siberian tigers simultaneously photographed within any given time period are calculated. Its mathematical expression is as follows:

[0039]

[0040] in, This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. The image of the Siberian tiger individual and the right side of the Siberian tiger individual is numbered in the image database. The distance cost of an individual Siberian tiger This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. The first Siberian tiger individual Zhou was numbered The wildlife conservation camera captured the image. This indicates that the image of the right side of an individual Siberian tiger, numbered [number missing], is located in the image database. The first Siberian tiger individual Zhou was numbered The wildlife conservation camera captured the image. Indicates the number is Wildlife camera and number The distance of the wildlife camera;

[0041] The cost matrix is ​​calculated based on the distance cost set, and its mathematical expression is as follows:

[0042]

[0043] in, This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. The image of the Siberian tiger individual and the right side of the Siberian tiger individual is numbered in the image database. The matching cost of individual Siberian tigers; This indicates the distance between the two farthest wildlife cameras;

[0044] Constructing an extended cost matrix , Its mathematical expression is as follows:

[0045] .

[0046] Furthermore, the specific implementation process of step S4.5 is as follows:

[0047] Based on the output of the Hungarian matching algorithm, for each matching pair ,if and Then, the image database containing the left side image of the Siberian tiger will be numbered as follows: The image of the Siberian tiger individual and the right side of the Siberian tiger individual is numbered in the image database. The Siberian tiger individuals are associated with the same Siberian tiger individual; if and This indicates that the image number in the database for the left side of the Siberian tiger is [number missing]. No corresponding image number for the Siberian tiger individual was found in the Siberian tiger right side body image database. Individual Siberian tigers; if and This indicates that the image number in the database for the right side of the Siberian tiger is [number missing]. No corresponding image number for the Siberian tiger individual was found in the Siberian tiger left side body image database. An individual Siberian tiger.

[0048] The beneficial effects of this invention are:

[0049] (1) Improved detection accuracy: In this invention, the Siberian tiger individual after the left and right side images are associated has bilateral information. Even if only one side is photographed in the wild, researchers can accurately match the correct Siberian tiger individual, thus improving the detection accuracy.

[0050] (2) In this invention, by linking the left and right side images of an individual Siberian tiger, the spatiotemporal information of the Siberian tiger is fully utilized, which can provide a solid data foundation for population dynamic monitoring. Researchers can more accurately estimate the population size and calculate the home range.

[0051] (3) This invention does not rely on human expert experience, shortens the detection time, and enables multi-person collaborative detection operation, making it convenient to operate. Attached Figure Description

[0052] Figure 1 The flowchart illustrates a method for associating left and right side images of an individual Siberian tiger based on spatiotemporal information, as provided by this invention.

[0053] Figure 2 A diagram illustrating key information about the Siberian tiger. Detailed Implementation

[0054] This invention provides a method for associating left and right side images of a Siberian tiger individual based on spatiotemporal information. First, data of Siberian tiger individuals captured by wildlife cameras are collected and labeled. Then, key point detection algorithms are used to predict the key point information and body orientation of the Siberian tiger individual. Individual information of the Siberian tiger is estimated based on an individual recognition algorithm. Finally, spatiotemporal information is combined to associate Siberian tiger individuals in the left and right side image database.

[0055] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0056] See Figure 1 As shown, the present invention provides a method for associating left and right side images of an individual Siberian tiger based on spatiotemporal information. The specific implementation process is as follows:

[0057] Step S1: Collect and label data on individual Siberian tigers captured by wildlife cameras;

[0058] S1.1: Collect images of individual Siberian tigers captured by wildlife conservation cameras.

[0059] S1.2: Label the location and key information of individual Siberian tigers in the images, and assign individual numbers.

[0060] S1.3: Statistics on the time information of individual Siberian tiger images being captured and the corresponding wildlife protection camera numbers.

[0061] S1.4: Collect the latitude, longitude, and attitude information of the installation location of the field protection camera.

[0062] Step S2: Predict key point information and body orientation of an individual Siberian tiger based on a key point detection algorithm;

[0063] S2.1: Based on the location and key point information of the Siberian tiger individuals marked in step S1.2, construct a Siberian tiger key point detection dataset.

[0064] S2.2: Using a keypoint detector (a neural network-based keypoint detector, including but not limited to PoseNet, HRNet, yolov8-pose, etc.), the Siberian tiger keypoint detection dataset is used to predict the side profile image of an individual Siberian tiger, obtaining the location and keypoint information of the individual tiger. Specifically, such as... Figure 2 As shown, the location information is the information of the bounding box containing the region of the Siberian tiger individual; there are six key points: key point 1 is the connection between the head and body of the Siberian tiger individual; key point 2 is the connection between the rump and tail of the Siberian tiger individual; key points 3 to 6 are the connections between the left forelimb, right forelimb, left hindlimb, and right hindlimb of the Siberian tiger individual and the body, respectively. By comparing the x-coordinates of key point 1 and key point 2, the side orientation of the Siberian tiger individual in the side-view image is determined. If key point 1 is to the left of key point 2, the side-view image of the Siberian tiger individual is the left side of the Siberian tiger individual; if key point 1 is to the right of key point 2, the side-view image of the Siberian tiger individual is the right side of the Siberian tiger individual.

[0065] S2.3: Construct a database of images of the left side of a Siberian tiger and an image database of images of the right side of a Siberian tiger based on the images of the side of the tiger predicted in step S2.2.

[0066] Step S3: Estimate individual information of Siberian tigers based on individual identification algorithms;

[0067] S3.1: First, based on the location and key point information of the Siberian tiger individual marked in step S1.2, the torso region image is extracted from the left side body image of the Siberian tiger individual, and then the torso region image is associated with the individual number marked in step S1.2.

[0068] S3.2: Assume that the database contains images of the left side of a Siberian tiger. Only, individual number is The database contains images of the right side of an individual Siberian tiger. Only, individual number is By horizontally flipping the right-side body image from the Siberian tiger individual image database and merging it with the left-side body image from the Siberian tiger individual image database, the desired result can be obtained. Images from the merged image database, representing individual Siberian tigers, were used as the training dataset for Siberian tiger identification.

[0069] S3.3: Construct an individual recognition network model (a classification model based on neural networks, including but not limited to ResNet, VIT, etc.). Use this individual recognition network model to extract features of the Siberian tiger's lateral trunk region image on the Siberian tiger individual recognition training dataset. Then, through the classification module (the classification module is part of the individual recognition network model, which is the last classification layer of the network, including fully connected layers and softmax layers: generally designed so that after feature output, a fully connected layer is connected, and then a softmax layer is connected to output the confidence score of each individual) to predict the Siberian tiger individual information.

[0070] Step S4: Combine spatiotemporal information to associate Siberian tiger individuals in the left and right side image databases;

[0071] S4.1: Associate the captured images of individual Siberian tigers with the individual ID information marked in step S1.2 and the wildlife camera IDs and time information captured in step S1.3 to construct a Siberian tiger spatiotemporal information database.

[0072] S4.2: Obtain the latitude and longitude information of the wildlife camera based on the positioning module inside the wildlife camera, and calculate the relative distance matrix of the wildlife camera based on the latitude and longitude information. ,in For the number of wildlife conservation cameras, Indicates the number is Wildlife camera and number The distance of the wildlife protection camera.

[0073] S4.3: Constructing an individual spatiotemporal matrix of the left side body image of an individual Siberian tiger based on the Siberian tiger spatiotemporal information database. Individual spatiotemporal matrix of the right side body image of an individual Siberian tiger ,in, This represents the number of Siberian tiger individuals in the database containing images of the left side of their bodies. This represents the number of Siberian tiger individuals in the database containing images of the right side of their bodies. The number of weeks the wildlife camera worked. This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. Was the Siberian tiger individual numbered in week k? The value is 0 when the wildlife protection camera captures the image and 1 when the image is captured. This indicates that the image of the right side of an individual Siberian tiger, numbered [number missing], is located in the image database. The first Siberian tiger individual Is Zhou numbered? The value is 0 when the wildlife camera captures the image and 1 when the image is captured.

[0074] S4.4: Constructing the cost matrix based on the individual spatiotemporal matrix of the left side image of an individual Siberian tiger, the individual spatiotemporal matrix of the right side image of an individual Siberian tiger, and the relative distance matrix of the wildlife camera. The lower the cost, the greater the likelihood that the two are the same Siberian tiger individual.

[0075] Specifically, the spatiotemporal matrices of the left and right body images of individual Siberian tigers are traversed along the time dimension. At weekly intervals, the distance cost sets of both left and right Siberian tigers that were simultaneously photographed within all time periods are calculated. Its mathematical expression is as follows:

[0076]

[0077] in, This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. The image of the Siberian tiger individual and the right side of the Siberian tiger individual is numbered in the image database. The distance cost of an individual Siberian tiger This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. The first Siberian tiger individual Zhou was numbered The wildlife conservation camera captured the image. This indicates that the image of the right side of an individual Siberian tiger, numbered [number missing], is located in the image database. The first Siberian tiger individual Zhou was numbered The wildlife conservation camera captured the image. Indicates the number is Wildlife camera and number The distance of the wildlife camera; This indicates the number of weeks the wildlife camera has been in operation.

[0078] The cost matrix is ​​calculated based on the distance cost set, and its mathematical expression is as follows:

[0079]

[0080] in, This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. The image of the Siberian tiger individual and the right side of the Siberian tiger individual is numbered in the image database. The matching cost of individual Siberian tigers; This represents the distance between the two furthest wildlife cameras. If individual Siberian tiger i on the left and individual Siberian tiger j on the right were present in the same week, then... = If individual Siberian tiger i on the left and individual Siberian tiger j on the right did not appear in the same week, then = .

[0081] Constructing an extended cost matrix , Its mathematical expression is as follows:

[0082]

[0083] Among them, if ,but In other cases, .

[0084] S4.5: Solve the extended cost matrix using the Hungarian matching algorithm;

[0085] Based on the output of the Hungarian matching algorithm, for each matching pair ,if and Then, the image database containing the left side image of the Siberian tiger will be numbered as follows: The image of the Siberian tiger individual and the right side of the Siberian tiger individual is numbered in the image database. The Siberian tiger individuals are associated with the same Siberian tiger individual; if and This indicates that the image number in the database for the left side of the Siberian tiger is [number missing]. No corresponding image number for the Siberian tiger individual was found in the Siberian tiger right side body image database. Individual Siberian tigers; if and This indicates that the image number in the database for the right side of the Siberian tiger is [number missing]. No corresponding image number for the Siberian tiger individual was found in the Siberian tiger left side body image database. An individual Siberian tiger.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. However, these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for associating left and right images of a Siberian tiger individual based on spatio-temporal information, characterized in that, The method comprises the following steps: Step S1: Collect and label data of a Siberian tiger individual photographed by a wildlife camera; Step S2: Predict key point information and body side orientation of the Siberian tiger individual based on a key point detection algorithm; Step S3: Estimate individual information of the Siberian tiger based on an individual recognition algorithm; Step S4: Correlate the Siberian tiger individual in left and right body side image libraries based on spatio-temporal information.

2. The method according to claim 1, wherein, The specific implementation process of step S1 is as follows: S1.1: Collect images of a Siberian tiger individual photographed by a wildlife camera; S1.2: Label the position and key point information of the Siberian tiger individual in the Siberian tiger individual image, and set an individual number; S1.3: Statistically analyze the time information and corresponding wildlife camera number of the Siberian tiger individual image; S1.4: Statistically analyze the latitude and longitude information and attitude information of the installation position of the wildlife camera.

3. The method according to claim 1, wherein, The specific implementation process of step S2 is as follows: S2.1: Based on the labeled position and key point information of the Siberian tiger individual, a Siberian tiger key point detection data set is constructed; S2.2: Based on the key point detector, the Siberian tiger key point detection data set is used to predict the body side image of the Siberian tiger individual, so as to obtain the position and key point information of the Siberian tiger individual; S2.3: According to the predicted body side image of the Siberian tiger individual, a left body image library of the Siberian tiger individual and a right body image library of the Siberian tiger individual are constructed.

4. The method according to claim 3, wherein, The labeled position information of the Siberian tiger individual is the information of a frame containing the Siberian tiger individual region.

5. The method according to claim 3, wherein, The labeled key points of the Siberian tiger individual include six key points: the first key point is the connection between the head and the body of the Siberian tiger individual; the second key point is the connection between the buttocks and the tail of the Siberian tiger individual; The third to sixth key points are the connections between the left forelimb, right forelimb, left hind limb and right hind limb of the Siberian tiger individual and the body.

6. The method of claim 5, wherein the method comprises: By comparing the horizontal coordinates of the first key point and the second key point, the body side orientation of the Siberian tiger individual in the body side image of the Siberian tiger individual is determined; if the first key point is located on the left side of the second key point, the body side image of the Siberian tiger individual is the left body image of the Siberian tiger individual; if the first key point is located on the right side of the second key point, the body side image of the Siberian tiger individual is the right body image of the Siberian tiger individual.

7. The method according to claim 1, wherein, The specific implementation process of step S3 is as follows: S3.1: Based on the labeled position and key point information of the Siberian tiger individual, a trunk region image is cut out from the left body image of the Siberian tiger individual, and the trunk region image is associated with the labeled individual number; S3.2: The right body image in the right body image library of the Siberian tiger individual is combined with the left body image in the left body image library of the Siberian tiger individual after being horizontally flipped, and the images in the combined image library are used as a Siberian tiger individual recognition training data set; S3.3: A individual recognition network model is constructed, features of the Siberian tiger body side trunk region image are extracted on the Siberian tiger individual recognition training data set, and individual information of the Siberian tiger is predicted through a classification module.

8. The method of claim 1, wherein the method is based on spatio-temporal information of the Siberian tiger. The specific implementation process of step S4 is as follows: S4.1: The photographed Siberian tiger individual image is associated with the labeled individual number information, the wildlife camera number and the photographed time information in step S1.3, and a Siberian tiger spatio-temporal information library is constructed; S4.2: Obtain the latitude and longitude information of the wildlife camera based on the positioning module inside the wildlife camera, and calculate the relative distance matrix of the wildlife camera according to the latitude and longitude information , is the number of wildlife cameras, represents the distance between the wildlife camera numbered and the wildlife camera numbered ; S4.3: Constructing an individual spatiotemporal matrix of the left side body image of an individual Siberian tiger based on the Siberian tiger spatiotemporal information database. Individual spatiotemporal matrix of the right side body image of an individual Siberian tiger , This represents the number of Siberian tiger individuals in the database containing images of the left side of their bodies. This represents the number of Siberian tiger individuals in the database containing images of the right side of their bodies. The number of weeks the wildlife camera worked. This indicates that the image of the left side of an individual Siberian tiger, numbered [number missing], is located in the image database. Was the Siberian tiger individual numbered in week k? The value is 0 when the wildlife protection camera captures the image and 1 when the image is captured. This indicates that the image of the right side of an individual Siberian tiger, numbered [number missing], is located in the image database. The first Siberian tiger individual Is Zhou numbered? The value is 0 when the wildlife protection camera captures the image and 1 when the image is captured. S4.4: constructing a cost matrix based on the individual spatio-temporal matrix of the left side body image of the individual Siberian tiger, the individual spatio-temporal matrix of the right side body image of the individual Siberian tiger, and the relative distance matrix of the wildlife camera ; An extended cost matrix is constructed; S4.5: Apply the Hungarian matching algorithm to solve the extended cost matrix.

9. The method according to claim 8, wherein, The specific implementation process of step S4.4 is as follows: The individual space-time matrix of the left body image of the Siberian tiger individual and the individual space-time matrix of the right body image of the Siberian tiger individual are traversed in the time dimension, and the distance cost set of the left Siberian tiger individual and the right Siberian tiger individual that are simultaneously photographed in all time is counted at intervals of a week The mathematical expression is as follows: ； wherein, denotes the distance cost between the Siberian tiger individual numbered in the Siberian tiger individual left body image library and the Siberian tiger individual numbered in the Siberian tiger individual right body image library, denotes the Siberian tiger individual numbered in the Siberian tiger individual left body image library was photographed by the wildlife camera numbered in the 1th week; denotes the Siberian tiger individual numbered in the Siberian tiger individual right body image library was photographed by the wildlife camera numbered in the 1th week; denotes the distance between the wildlife camera numbered and the wildlife camera numbered .​​ According to the distance cost set, the cost matrix is calculated, and the mathematical expression is as follows: ； wherein, represents the matching cost between the left body image of the Siberian tiger individual numbered in the left body image library of Siberian tiger individuals and the right body image of the Siberian tiger individual numbered in the right body image library of Siberian tiger individuals; represents the distance value between the two farthest wildlife cameras. Constructing an extended cost matrix , The mathematical expression is as follows: 。 10. The method according to claim 9, wherein, The specific implementation process of step S4.5 is as follows: According to the output of the Hungarian matching algorithm, for each matching pair if and , then the Siberian tiger individual numbered in the Siberian tiger individual left body image library is associated with the Siberian tiger individual numbered in the Siberian tiger individual right body image library as the same Siberian tiger individual; if and , then it indicates that the Siberian tiger individual numbered in the Siberian tiger individual left body image library does not find the corresponding Siberian tiger individual numbered in the Siberian tiger individual right body image library; if and , then it indicates that the Siberian tiger individual numbered in the Siberian tiger individual right body image library does not find the corresponding Siberian tiger individual numbered in the Siberian tiger individual left body image library.

Images