Wild animal abnormal activity track real-time tracking method and system

Abstract

The invention relates to the technical field of trajectory tracking, in particular to a wild animal abnormal activity trajectory real-time tracking method and system. The method comprises the following steps: acquiring position information of wild animals; constructing predicted position information based on the state transition matrix, and constructing a covariance matrix based on the predicted position information and the position information, thereby determining a measurement noise covariance matrix; constructing an acquisition window, calculating a deviation degree of each dimension based on the acquisition window, determining an abnormal weight in combination with a difference between a measured value and a predicted value, further determining a diagonal residual matrix, and then obtaining a Kalman gain matrix in combination with a measurement noise covariance matrix; and filtering and correcting the output signal based on the Kalman gain matrix, and monitoring an abnormal track area after generating a heat map from the signal. According to the invention, more accurate tracking monitoring is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of trajectory tracking, in particular to a wild animal abnormal activity trajectory real-time tracking method and system. BACKGROUND

[0002] With the situation of human poaching, climate change and the like, the animal habitats in nature are being destroyed, and many wild animals such as giant pandas and South China tigers are gradually facing extinction. At the same time, there are some wild animals that are injured or illegally traded and need to be treated or helped in time. Therefore, in order to protect and study the above-mentioned wild animals, many wild animal protection bases and zoos have been built. The wild animal protection bases and zoos provide a refuge for endangered wild animals from habitat destruction, poaching or climate change threats, delay species extinction through scientific breeding programs (such as captive breeding, genetic diversity management), and release healthy animal populations into the wild to help species diversity. In the real-time detection and analysis of the activity trajectory of endangered animals in wild animal protection bases and the like, the daily trajectory of the animals is usually detected and tracked in real time to detect the stability of the animal's living state. When the animal has an abnormal activity trajectory due to some special circumstances and deviates from the daily activity range, it may have an impact on the new area, causing the wild protected animals to encounter danger.

[0003] In order to timely handle the state changes of animals and assist the protection and research of wild animal protection bases and the like, real-time automatic tracking of animal abnormal activity trajectories by using tracking devices is an important research direction at present. At present, the abnormal activity trajectory of the animal is mainly tracked and monitored by using a positioning tracking device. However, due to the interference of the facilities in the living scene of the animal, such as the existence of dense forests, gorges or mountains in the corresponding protection base, the signal is blocked, and signal attenuation, sudden abnormal signal and the like occur when the positioning signal of the animal is acquired. There is a big difference between the measurement noise covariance matrix preset by the conventional Kalman filter and the interference noise actually received during signal acquisition, which further causes the animal activity trajectory line to deviate abnormally, and the problem still needs to be further solved. SUMMARY

[0004] In order to solve the technical problem of poor measurement accuracy caused by noise, the present application provides a wild animal abnormal activity trajectory real-time tracking method and system, and the technical solution adopted is as follows: In a first aspect, the present application provides a wild animal abnormal activity trajectory real-time tracking method, which comprises the following steps: Collecting the position information of the wild animal; pre-set a transition state matrix through speed change of the wild animal; construct the predicted position information of the current collection based on the transition state matrix and the predicted position information collected last time; construct a position covariance matrix based on the vector difference between the position information of the current collection and the predicted position information; construct a measurement noise covariance matrix of the current collection through the position covariance matrix of the current collection and the measurement noise covariance matrix collected last time; construct a collection window for a pre-set number of collections before each collection, calculate the deviation degree of each dimension data through the collection window; determine the abnormal weight on each dimension based on the difference between the measured value of each dimension and the predicted value of each dimension and the deviation degree; construct a diagonal residual matrix of the abnormal weight of all dimensions; reconstruct the Kalman gain of Kalman filtering based on the measurement noise covariance matrix and the diagonal residual matrix, and obtain a Kalman gain matrix; perform filtering correction on the output signal based on the Kalman gain matrix, and monitor the abnormal trajectory area after generating a heat map of the signal.

[0005] In the above scheme, the adaptive Kalman gain is constructed, and when the animal moves to the dense forest and the like to cause shielding, the positioning signal may cause signal attenuation, mutation and abnormal signal and the like. The conventional Kalman filtering uses a fixed measurement noise covariance matrix, and when the shielding occurs in the environment during the animal activity, the signal obtained by the tracking device is attenuated, mutated and abnormal, and there is a large difference between the pre-set measurement noise covariance matrix and the interference noise received during actual signal collection, which may cause large trajectory deviation. The adaptive measurement noise covariance matrix is constructed, the weight of the measured data is reduced when the signal is attenuated due to shielding, and the diagonal residual matrix is used to suppress the mutation abnormal value when the single dimension variable is mutated, so as to avoid the influence of the mutation of a single dimension on all dimensions when the algorithm corrects the signal, and the optimized Kalman filtering is further used to realize more accurate tracking and monitoring through accurate correction.

[0006] In one embodiment, the dimensions of the position information include longitude, latitude, longitude speed, latitude speed, longitude acceleration and latitude acceleration.

[0007] In one embodiment, the method of pre-setting the transition state matrix through the speed change of the wild animal is: select the first n data of the position information collected each time; When the total speed of the previous n data continuously increases, the uniform acceleration mode is set; when the total speed of the previous n data continuously decreases, the uniform deceleration mode is set; when the total speed of the previous n data is unchanged, the uniform speed mode is set, and then a state transition matrix is constructed through a motion model; the total speed is the sum of the longitude speed and the latitude speed.

[0008] In one embodiment, the method for constructing the position covariance matrix is: , the position vector representing the position information collected for the kth time, the position vector representing the predicted position information collected for the kth time, represents the transpose operation of a matrix, the position covariance matrix collected for the kth time; wherein the position vector is a vector composed of all dimensions of the position information.

[0009] In one embodiment, the method for constructing the measurement noise covariance matrix is: , the smoothing factor, the measurement noise covariance matrix collected for the (k-1)th time, the position covariance matrix collected for the kth time, the diagonal matrix function, the measurement noise covariance matrix collected for the kth time.

[0010] In one embodiment, the method for calculating the deviation degree of the data of each dimension through the collection window is: For the data of each dimension in the collection window, after fitting the data of each dimension, the residual of the fitted value and the measured value is calculated, and the standard deviation of all residuals is taken as the deviation degree of the data of each dimension.

[0011] In one embodiment, the abnormal weight is positively correlated with the difference between the measured value and the predicted value, and is negatively correlated with the deviation degree; the predicted value of each dimension is the value corresponding to the dimension of the position vector of the predicted position information.

[0012] In one embodiment, the diagonal residual matrix is a diagonal matrix composed of the abnormal weights of all dimensions in the order of the position vector.

[0013] In one embodiment, the expression of the Kalman gain matrix is: , the prior error covariance matrix collected for the uth time, the longitude and latitude observation matrix, denotes a transpose matrix of the latitude and longitude observation matrix, denotes a diagonal residual matrix of the u-th acquisition, denotes a measurement noise covariance matrix of the u-th acquisition, denotes a Kalman gain matrix of the u-th acquisition; the latitude and longitude observation matrix is set to 1 for the values of longitude and latitude and 0 for the rest of the dimensions.

[0014] In a second aspect, the embodiments of the present application further provide a wild animal abnormal activity track real-time tracking system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of the wild animal abnormal activity track real-time tracking method in any of the above aspects when executing the computer program.

[0015] The present application has the following beneficial effects: The present application constructs an adaptive Kalman gain, considers that when the animal moves to the dense forest and the like to generate shielding, the positioning signal may cause signal attenuation, mutation abnormal signal and the like. The conventional Kalman filter uses a fixed measurement noise covariance matrix, and the signal obtained by the tracking device is affected by the shielding in the environment when the animal moves, resulting in signal attenuation, mutation abnormality and the like. There is a large difference between the preset measurement noise covariance matrix and the interference noise actually received during signal acquisition, which may further cause a large track deviation. The present application constructs an adaptive measurement noise covariance matrix, reduces the weight of the measured data when the signal attenuation occurs due to shielding, and suppresses the mutation abnormal value when a single dimension variable mutation occurs through the diagonal residual matrix, so as to avoid the algorithm from being affected by the single dimension mutation when correcting the signal, and further use the optimized Kalman filter to realize more accurate tracking and monitoring through accurate correction. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0017] Figure 1 A wild animal abnormal activity track real-time tracking method flowchart provided by an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined object of the application, the specific implementation, structure, features and effects of the wild animal abnormal activity trajectory real-time tracking method and system according to the present application are described in detail as follows in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0020] A wild animal abnormal activity trajectory real-time tracking method and system embodiment: The specific scheme of the wild animal abnormal activity trajectory real-time tracking method and system provided by the present application is described in detail below in combination with the drawings.

[0021] Please refer to Figure 1 which shows a wild animal abnormal activity trajectory real-time tracking method flowchart provided by an embodiment of the present application, the method includes the following steps: Step S001, collecting position information of wild animals.

[0022] In the real-time automatic tracking of animal abnormal activity trajectory in a wild animal protection base, the scene usually shows relatively complex scene features, and there are usually dense forests, canyons and mountain bodies and other characteristic scenes. Therefore, when positioning and tracking the movement trajectory of animals in the wild animal protection base, a positioning and tracking necklace is usually used to obtain real-time positioning information of the animals in the protection base. In this embodiment, wild animals in the wild animal protection base are selected as the tracking target of abnormal activity trajectory, and a wild animal tracker based on nRF9160 is configured on the wild animals to obtain the GPS positioning signal of the wild animals in real time, i.e. to obtain the position information. In this embodiment, the position information is represented by longitude, latitude, longitude speed, latitude speed, longitude acceleration and latitude acceleration.

[0023] At this point, the position information of each sampling is obtained.

[0024] Step S002, constructing predicted position information based on a state transition matrix, constructing a covariance matrix based on the predicted position information and the position information, and determining a measurement noise covariance matrix therefrom.

[0025] Due to the shielding conditions such as dense forest, canyon and mountain in the nature reserve scene or the rainstorm caused by weather change, the acquired signal is disturbed by complex noise. In order to ensure the stability and effectiveness of the GPS positioning signal acquired by the wild animals in the nature reserve when moving, Kalman filtering is selected to predict and correct the positioning signal of the activity track of the wild animals in the nature reserve scene.

[0026] When the conventional Kalman filtering is used to predict and correct the positioning signal of the activity track of the wild animals in the nature reserve scene, considering that the wild animals are affected by the natural environment when moving in the nature reserve, the actual generated interference noise is real-time change. For example, when the rainstorm occurs, the noise suddenly increases, and when passing through the dense forest, the interference of leaves, tree trunks and other obstacles occurs. When the above influences occur, the collected position information is affected, and there is a large deviation from the predicted position information.

[0027] The data of the predicted position information is obtained by multiplying the state transition matrix and the data of the predicted position information of the last time. The state transition matrix is determined by the motion model. The motion model is constructed by selecting the speed change of the signal acquisition of the previous n times corresponding to each collection time. In order to increase the prediction accuracy, the value range of n is set to [4, 7], and n is taken as the experience value 5 in this embodiment. According to the speed data change, when the speed increases in the order of time, that is, the speed at each time is greater than the speed at the previous time, the uniform acceleration mode is set. When the speed decreases in the order of time, that is, the speed at each time is less than the speed at the previous time, the uniform deceleration mode is set, and other conditions are set to the uniform speed mode. The transition state matrix is obtained based on different modes. The predicted position information of the first time is the position information of the first time.

[0028] For each sampling, the position information is formed into a position vector, and the position vector is a vector formed by all dimensions of the position information. The predicted position information is also formed into a position vector, and the position covariance matrix is formed based on the difference between the two position vectors and its transpose.

[0029] Preferably, in this embodiment, the position covariance matrix is constructed by: , The position vector of the position information collected for the kth time is represented by The position vector of the predicted position information collected for the kth time is represented by The transpose operation of the matrix is represented by The position covariance matrix collected for the kth time is represented by. The diagonal data of the obtained covariance matrix is the square of the deviation of the data of the corresponding dimensional position information and the data of the predicted position information, which represents the influence of the signal after being disturbed by the interference noise.

[0030] When the wild animals being tracked are affected by the environment (e.g., the signal is scattered due to rain when the animals are running in the rain), the acquired position information deviates greatly from the predicted position information, and the diagonal values of the covariance matrix increase significantly; when the animals move in a normal, non-interfered environment, the diagonal values of the acquired covariance matrix are relatively small.

[0031] The conventional Kalman filter usually processes the preset measurement noise covariance matrix when predicting and correcting the signal; however, in the case of signal scattering caused by sudden rainstorm or forest shelter, there is a large difference between the preset measurement noise covariance matrix and the interference noise during actual position information acquisition, which leads to a decline in the filtering effect.

[0032] Therefore, the measurement noise covariance matrix acquired this time is obtained by weighting the position covariance matrix acquired each time and the measurement noise covariance matrix acquired last time.

[0033] Preferably, in this embodiment, the expression of the measurement noise covariance matrix is: , represents a smoothing factor, represents the measurement noise covariance matrix acquired at the k-1th time, represents the position covariance matrix acquired at the kth time, represents a diagonal matrix function, represents the measurement noise covariance matrix acquired at the kth time.

[0034] The diagonal matrix function means that the elements outside the main diagonal are all set to 0; the measurement noise covariance matrix acquired at the first time is set by the error of the measuring device itself.

[0035] When the positioning signal of the wild animals being tracked is affected by the environment and produces errors, the diagonal function of the position covariance matrix increases significantly, and the updated measurement noise covariance matrix also increases accordingly; when the animals move in a normal, non-interfered environment, the residual covariance term acquired is extremely small, and the noise covariance matrix gradually tends to be stable with updating.

[0036] At this point, the measurement noise covariance matrix is obtained.

[0037] Step S003: A collection window is constructed, the deviation degree of each dimension is calculated based on the collection window, the abnormal weight is determined by combining the difference between the measured value and the predicted value, the diagonal residual matrix is further determined, and then the Kalman gain matrix is obtained in combination with the measurement noise covariance matrix.

[0038] Considering that the position information of the animal is located by using the tracking device, the positioning signal of the animal passing through the forest is disturbed by the obstacles in all directions. Since the latitude, longitude, speed and acceleration obtained by the tracking device are usually obtained by multiple sensors, when collecting and obtaining the signals of multiple dimensions synchronously, such as latitude and longitude signals, instantaneous signal loss or jump may occur due to the animal being blocked or electromagnetic interference of thunderstorm weather, resulting in instantaneous abnormal values, and thus the reliability of the measured data is relatively lower.

[0039] Based on the above analysis, in order to deal with the possible abnormal values, the weight of the corresponding dimension of the measurement noise covariance matrix is calculated. For each collection, the first preset number M of each sampling is used as a sliding window. In order to ensure that the obtained data will not be too large due to the time distance, the value range of M is set to [3, 6], and the value of M in this embodiment is 5.

[0040] For the case where the number of previous collections is less than M, the data is supplemented in a mean filling manner to construct a noise weight index for representing the increased weight of the corresponding dimension of the measurement noise covariance matrix when the abnormal value occurs in the corresponding dimension.

[0041] In the Kalman filter, an observation matrix is set, wherein each dimension corresponds to an observation matrix, and the observation matrix of each dimension is 1 for the value in the dimension and 0 for the data in the remaining dimensions.

[0042] For each sliding window, the data in each dimension is fitted to calculate the residual, and then the standard deviation of the residual is calculated, which represents the deviation degree of the measured value and the predicted value.

[0043] Based on the comparison of the difference and deviation degree between the measured value of each dimension and the predicted value of each dimension, the abnormal weight on each dimension is determined.

[0044] The abnormal weight is positively correlated with the difference between the measured value and the predicted value, and is negatively correlated with the deviation degree.

[0045] It should be noted that the positive correlation means that when one variable increases, the other variable also increases, the two variables change in the same direction, and when one variable changes from large to small or from small to large, the other variable also changes from large to small or from small to large. The specific relationship is determined by actual application, and the present application does not make special limitation.

[0046] It should be noted that the negative correlation means that when one variable increases, the other variable decreases, and the two variables change in opposite directions, and when one variable changes from large to small or from small to large, the other variable also changes from small to large or from large to small. The specific relationship is determined by actual application, and the present application does not make special limitation.

[0047] Preferably, in the embodiment, the expression of the abnormal weight is: , represents the observation matrix of dimension c, represents the data of dimension c of the position information of the kth acquisition, represents the position vector of the predicted position information of the kth acquisition, represents the residual standard deviation of the data of dimension c in the corresponding sliding window at the kth acquisition, represents a very small positive number, used to ensure that the denominator is not 0, represents a normalization function, represents the abnormal weight of dimension c of the kth acquisition.

[0048] All the abnormal weights of all dimensions of the kth acquisition are arranged in the order of the position vector to form a diagonal matrix, which is denoted as the diagonal residual matrix of the noise weight.

[0049] The weights calculated for the corresponding dimensions form a diagonal matrix, when there are abnormal values such as jumps and losses in the data of the corresponding dimensions, the value calculated is much higher than the residual standard deviation in the sliding window, the D value obtained by calculation is higher, and the weight obtained is larger; when the data is calculated under normal and undisturbed conditions, the value calculated is smaller, the D value obtained by calculation is smaller, and the weight obtained is smaller.

[0050] Based on the measurement noise covariance matrix and the diagonal residual matrix, the Kalman gain of the Kalman filter is reconstructed to obtain the Kalman gain matrix.

[0051] Preferably, the expression of the Kalman gain matrix is: , represents the prior error covariance matrix of the uth acquisition, represents the longitude and latitude observation matrix, represents the transpose matrix of the longitude and latitude observation matrix, represents the diagonal residual matrix of the uth acquisition, represents the measurement noise covariance matrix of the uth acquisition, represents the Kalman gain matrix of the uth acquisition.

[0052] The longitude and latitude observation matrix is set to 1 for the values of longitude and latitude, and 0 for the remaining dimensions; the prior error covariance matrix is obtained by adding the process noise covariance matrix to the state transition matrix and the error covariance matrix at the previous time; the process noise covariance matrix is obtained by deducing the motion model. The first error covariance matrix is obtained by the error of the device itself.

[0053] At this point, the Kalman gain matrix of the Kalman filter is obtained.

[0054] In step S004, the output signal is corrected based on the Kalman gain matrix, and the signal is generated into a heat map to monitor the abnormal trajectory area.

[0055] The positioning signal of the animal in the wildlife protection area obtained by configuring the positioning device is filtered and corrected by the Kalman filter optimized by the Kalman gain matrix, and the filtered positioning signal data is output. The Kalman filter is a known technology, and the specific operation process will not be described in detail. The specified activity range in the wildlife protection area is obtained, the positioning signal data obtained is summarized into the corresponding area image, the motion trajectory of the animal in the past week is taken as a period, the trajectory of the animal in the period is summarized into the local area image, and then a heat map is generated according to the trajectory and by Folium. The heat map reflects the position preference of the animal in the wildlife protection area. The area with the least positioning information in the period is determined as the abnormal trajectory area (for example, a tiger usually patrols and captures prey in its own territory, and generally does not go to other tiger territories. When the target tiger goes to other tiger territories, it means that the target tiger has abnormal activity trajectory, which may cause a fight between different tigers). When the animal appears in the corresponding abnormal trajectory area, real-time trajectory tracking is performed on the animal when the abnormal activity trajectory appears by using the positioning device, and the abnormal activity trajectory tracking data is fed back in real time to warn, and real-time tracking of the abnormal activity trajectory of the animal is completed.

[0056] At this point, the real-time tracking of the abnormal activity trajectory of the animal is completed.

[0057] Based on the same inventive concept as the above method, the embodiments of the present application also provide a wildlife abnormal activity trajectory real-time tracking system, which comprises a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of any one of the above-mentioned wildlife abnormal activity trajectory real-time tracking methods are implemented.

[0058] It should be noted that the above-mentioned embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

[0059] The various embodiments described in this specification are presented by way of example, and each embodiment is not inherently more important than any other embodiment.

Claims

1. A method for real-time tracking of abnormal activity trajectory of wild animals, characterized in that, The method comprises the following steps: Collecting position information of wild animals; Pre-setting a transition state matrix based on speed changes of the wild animals; constructing predicted position information collected this time based on the transition state matrix and the last collected predicted position information; constructing a position covariance matrix based on vector differences between the collected position information and the predicted position information this time; and constructing a measurement noise covariance matrix collected this time based on the position covariance matrix collected this time and the measurement noise covariance matrix collected last time; For a preset number of times of collection before each time of collection, a collection window is constructed, and the deviation degree of each dimension data is calculated through the collection window; the abnormal weight on each dimension is determined based on the difference between the measured value of each dimension and the predicted value of each dimension and the deviation degree; a diagonal residual matrix is constructed based on the abnormal weights of all dimensions; and the Kalman gain of Kalman filtering is reconstructed based on the measurement noise covariance matrix and the diagonal residual matrix to obtain a Kalman gain matrix; Based on the Kalman gain matrix, the output signal is filtered and corrected, and the signal is monitored for an abnormal trajectory area after generating a heat map.

2. The method of claim 1, wherein the method comprises: The dimensions of the position information include longitude, latitude, longitude speed, latitude speed, longitude acceleration, and latitude acceleration.

3. The method of claim 1, wherein the method comprises: The method for pre-setting the transition state matrix based on the speed changes of the wild animals is: Selecting the first n times of data of each time of collecting position information; When the total speed of the first n times of data continuously increases, setting it as a uniform acceleration mode; when the total speed of the first n times of data continuously decreases, setting it as a uniform deceleration mode; and when the total speed of the first n times of data is unchanged, setting it as a uniform speed mode, and then constructing a state transition matrix through a motion model; the total speed is the sum of the longitude speed and the latitude speed.

4. The method of claim 1, wherein the method comprises: The construction method of the position covariance matrix is: position vector representing position information of the kth acquisition, position vector representing predicted position information of the kth acquisition, denotes a transposition operation of a matrix, denotes a position covariance matrix of the kth acquisition;​ The position vector is a vector composed of all dimensions of the position information.

5. The method of claim 1, wherein the method comprises: The construction method of the measurement noise covariance matrix is: , denotes a smoothing factor, denotes the measurement noise covariance matrix of the (k-1)th acquisition, denotes the position covariance matrix of the kth acquisition, denotes a diagonal matrix function, denotes the measurement noise covariance matrix of the kth acquisition.

6. The method of claim 1, wherein the method comprises: The method for calculating the deviation degree of each dimension data through the collection window is: For the data of each dimension in the collection window, the residual error between the fitting value and the measured value is calculated after fitting the data of each dimension, and the standard deviation of all residual errors is taken as the deviation degree of the data of each dimension.

7. The method of claim 1, wherein the method comprises: The abnormal weight is positively correlated with the difference between the measured value and the predicted value, and is negatively correlated with the deviation degree; and the predicted value of each dimension is the value of the corresponding dimension of the position vector of the predicted position information.

8. The method of claim 1, wherein the method comprises: The diagonal residual matrix is a diagonal matrix composed of the abnormal weights of all dimensions in the order of the position vector.

9. The method of claim 1, wherein the method comprises: The expression of the Kalman gain matrix is: , represents a prior error covariance matrix of the u-th acquisition, represents a latitude and longitude observation matrix, represents a transpose matrix of the latitude and longitude observation matrix, represents a diagonal residual matrix of the u-th acquisition, represents a measurement noise covariance matrix of the u-th acquisition, represents a Kalman gain matrix of the u-th acquisition; the latitude and longitude observation matrix is set to 1 for the values of longitude and latitude and 0 for the remaining dimensions.

10. A wild animal abnormal activity trajectory real-time tracking system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor implements the steps of the wild animal abnormal activity trajectory real-time tracking method according to any one of claims 1-9 when executing the computer program.

Images