Wild animal abnormal behavior intelligent analysis method and system

Abstract

The invention discloses an intelligent analysis method and system for abnormal behaviors of wild animals, and the method comprises the steps: recognizing the population type of the wild animals entering a monitoring region based on a static image collection system; on the basis of the wildlife population types, wildlife populations with abnormal trails are obtained; based on a static image acquisition system, obtaining a multidirectional side view of the wildlife population with the abnormal track; based on a dynamic image acquisition system, obtaining a top view of the wildlife population with the abnormal track; based on the top view of the wildlife population with abnormal trails, identifying wildlife individuals with abnormal trails; setting a part frame for the wild animal individual with the abnormal track, and obtaining a part frame change rate based on the part frame; and obtaining the abnormal behaviors and causes of the wild animals based on the part frame change rate. The problem that precision is insufficient due to the fact that track and individual behaviors cannot be analyzed at the same time in abnormal behavior analysis is solved. According to the invention, common analysis of the two abnormal behaviors is realized, and the analysis precision is improved.

Description

Technical Field

[0001] This application belongs to the field of wild animal behavior monitoring, and specifically, to an intelligent analysis method and system for abnormal behaviors of wild animals. Background Art

[0002] In wild animal protection, the analysis of abnormal behaviors of animals is an important task. Currently, the analysis of abnormalities in wild animals has begun based on image acquisition and processing technologies. However, in the current analysis of abnormal behaviors of wild animals, the main method is to set up monitoring devices on the migration tracks of wild animals to obtain the tracks of the animals, and further determine whether there are abnormal problems in the tracks of the animals. The analysis effect of individual abnormal behaviors of wild animals is poor. In addition, although the tracks of wild animals are relatively fixed in terms of destinations, their specific tracks may be variable, resulting in a low coupling between the already deployed monitoring system and the tracks, leading to poor tracking effects. In addition, continuously deploying monitoring devices in the track areas of wild animals will cause the complexity and construction cost of the entire system to continue to increase, which is not conducive to improving the robustness of the system and is not conducive to cost control.

[0003] Therefore, how to establish an intelligent analysis method for abnormal behaviors of wild animals that can reduce the construction cost of the system while ensuring high-precision analysis of the tracks and individual abnormal behaviors of wild animals at the same time is a technical problem that needs to be solved urgently by those skilled in the art. Summary of the Invention

[0004] In order to solve the problems in the prior art that the construction cost of the intelligent analysis system for abnormal behaviors of wild animals is too high, and at the same time the accuracy of the tracks and individual abnormal behaviors of wild animals is insufficient, this application discloses an intelligent analysis method and system for abnormal behaviors of wild animals, specifically: First aspect: An intelligent analysis method for abnormal behaviors of wild animals, the method includes: Based on a static image acquisition system, identify the types of wild animal populations entering the monitoring area; Based on the types of wild animal populations, obtain wild animal populations with abnormal tracks; Based on the static image acquisition system, obtain multi-directional side views of the wild animal populations with abnormal tracks; Based on a dynamic image acquisition system, obtain top views of the wild animal populations with abnormal tracks; Based on the top views of the wild animal populations with abnormal tracks, identify individual wild animals with abnormal tracks; Set part borders for the individual wild animals with abnormal tracks, and obtain the change rate of the part borders based on the part borders; Based on the change rate of the part border, obtain the abnormal behaviors and causes of wild animals.

[0005] Optionally, based on the static image acquisition system, identify the types of wild animal populations entering the monitoring area, including: Obtain the types of wild animal populations in the area and the normal distribution areas of the wild animal populations to obtain the normal monitoring areas; Based on the normal distribution areas, obtain the abnormal distribution areas of the wild animal populations; Based on the distribution spacing between the normal distribution areas and the abnormal distribution areas, obtain the abnormality degree of the distribution areas of the wild animal populations; Set up a static image acquisition system for the abnormal distribution areas to obtain abnormal monitoring areas; Obtain the entry status of wild animals in the normal monitoring areas and the abnormal monitoring areas, and identify the types of wild animal populations.

[0006] Optionally, based on the types of wild animal populations, obtain the wild animal populations with abnormal tracks, including: Set the types of wild animal populations that normally enter the normal monitoring areas to obtain the normal wild animal population types; Compare the types of wild animal populations entering the normal monitoring areas with the normal wild animal population types. If they are different, the wild animal populations entering the normal monitoring areas are the wild animal populations with abnormal tracks, and obtain the wild animal populations with abnormal tracks in the normal monitoring areas; Obtain the types of wild animal populations in the abnormal monitoring areas and obtain the wild animal populations with abnormal tracks in the abnormal monitoring areas; Both the wild animal populations with abnormal tracks in the normal monitoring areas and the wild animal populations with abnormal tracks in the abnormal monitoring areas are set as the wild animal populations with abnormal tracks; Obtain the types of wild animal populations with abnormal tracks and obtain the severity of the abnormal tracks of the wild animal populations with abnormal tracks.

[0007] Optionally, based on the static image acquisition system, obtain the multi-directional side views of the wild animal populations with abnormal tracks, including: Based on the static image acquisition system, obtain the multi-directional frame images of the wild animal populations of the tracks; Based on the multi-directional frame images and the orientations of all the image acquisition devices in the static image acquisition system, obtain the multi-directional side views.

[0008] Optionally, based on the dynamic image acquisition system, obtain the top views of the wild animal populations with abnormal tracks, including: After the static image acquisition system obtains the wildlife population with abnormal tracks, the dynamic image acquisition system moves towards the corresponding monitoring area; The dynamic image acquisition system reaches above the wildlife population with abnormal tracks and continuously obtains the top view of the wildlife population with abnormal tracks.

[0009] Optionally, identifying the individual wildlife with abnormal tracks based on the top view of the wildlife population with abnormal tracks includes: Based on the top view of the wildlife population with abnormal tracks, label the wildlife within the wildlife population with abnormal tracks to obtain individual labeling information; Based on the top view of the wildlife population with abnormal tracks, associate the individual labeling information with the top view of the individual wildlife with abnormal tracks to obtain the information of the individual wildlife with abnormal tracks.

[0010] Optionally, setting a part border for the individual wildlife with abnormal tracks and obtaining the change rate of the part border based on the part border includes: Obtain the body parts of the individual wildlife with abnormal tracks and set a side view part border on the multi-directional side view; Obtain the top view of the individual wildlife with abnormal tracks and set a top view part border based on the body parts; Perform coupling processing on the side view part border and the top view part border of the individual wildlife with abnormal tracks to obtain a part border; Obtain the change rate of all the side lines of the part border to obtain the change rate of the part border.

[0011] Optionally, obtaining the abnormal behaviors and causes of the wildlife based on the change rate of the part border includes: Obtain the occurrence times of the change rates of all the part borders of the individual wildlife with abnormal tracks to obtain the occurrence frequency of compulsive behaviors; Compare the occurrence frequency of compulsive behaviors with the preset occurrence frequency of compulsive behaviors. If the occurrence frequency of compulsive behaviors is not lower than the preset occurrence frequency of specific behaviors, determine that the individual wildlife with abnormal tracks has abnormal behaviors; Sort the occurrence frequencies of the change rates of the part borders of the individual wildlife with abnormal tracks that have abnormal behaviors to obtain high-frequency abnormal behaviors; Compare the high-frequency abnormal behaviors, the severity of the abnormal tracks of the wildlife population with abnormal tracks, and the causes of the abnormal behaviors of the wildlife to obtain the causes of the abnormal behaviors of the wildlife.

[0012] Second aspect: An intelligent analysis system for abnormal animal behavior, which is used to execute the intelligent analysis method for abnormal animal behavior as described in the first aspect, includes a static image acquisition system, a dynamic image acquisition system, an animal feature recognition system, a wild animal population recognition system with abnormal tracks, an individual wild animal recognition system with abnormal tracks, an individual part decomposition system of wild animals with abnormal tracks, an abnormal behavior analysis system and a database, and is characterized in that: The static image acquisition system is connected to the animal feature recognition system, and is used to obtain wild animal populations in the normal monitoring area and the abnormal monitoring area, and send the obtained monitoring information to the animal feature recognition system; The animal feature recognition system is connected to the wild animal population recognition system with abnormal tracks, and is used to obtain the features of wild animals in the monitoring area; The wild animal population recognition system with abnormal tracks is also connected to the dynamic image acquisition system, and is used to obtain the wild animal population information with abnormal tracks based on the wild animal features. After obtaining the wild animal population information with abnormal tracks, it sends a shooting instruction and monitoring information to the image acquisition system; The individual wild animal recognition system with abnormal tracks is simultaneously connected to the wild animal population recognition system with abnormal tracks and the dynamic image acquisition system, and is used to obtain the individual wild animal images with abnormal tracks; The individual part decomposition system of wild animals with abnormal tracks is connected to the individual wild animal recognition system with abnormal tracks, and is used to obtain the part frames of wild animals with abnormal tracks; The abnormal behavior analysis system is connected to the individual part decomposition system of wild animals with abnormal tracks and the wild animal population recognition system with abnormal tracks, and is used to obtain the types and causes of abnormal behaviors of the wild animal population with abnormal tracks and the individual wild animals with abnormal tracks; The abnormal behavior analysis system is also connected to the database, and the database is used to store the corresponding relationships between the change rates of the part frames of wild animals with abnormal tracks, the severity of abnormal tracks of wild animal populations with abnormal tracks, and the causes of abnormal behaviors of wild animals.

[0013] Optionally, the static image acquisition system further includes: The static image acquisition system includes a normal monitoring area and an abnormal monitoring area. The normal monitoring area is an existing image acquisition area, and the abnormal monitoring area is another image acquisition area delimited based on the existing image acquisition area; In the normal monitoring area and the abnormal monitoring area, a plurality of image acquisition devices are arranged on the edge line of the monitoring area.

[0014] The beneficial effects of this application include: 1. Reduced the implementation cost of analyzing abnormal behaviors of wild animals. In the technical solution of this application, a static and dynamic image acquisition system is established. In the static image acquisition system, the existing image acquisition systems are fully utilized. At the same time, based on the existing image acquisition systems, abnormal monitoring areas are selected, and corresponding image acquisition systems are established, realizing the effective application of existing resources. At the same time, the dynamic image acquisition system can reach the monitoring area by itself for image acquisition without fixed installation, enabling the dynamic image acquisition system to be recycled and reused, reducing the construction cost of the entire system and the implementation cost of the method.

[0015] 2. Broadened the scope of work for analyzing abnormal behaviors of wild animals. In the technical solution of this application, based on the static image acquisition system, wild animal populations with abnormal tracks are obtained. Then, based on the combined application of the static and dynamic image acquisition systems, the abnormal behaviors of individual wild animals with abnormal tracks are analyzed, thus realizing the combined analysis of abnormal behaviors of populations and individuals, and broadening the coverage scope of the work of analyzing abnormal behaviors.

[0016] 3. Reduced the resource consumption of the work of analyzing abnormal behaviors of wild animals. In the technical solution of this application, for the process of analyzing abnormal behaviors of wild animals, appearance analysis based on image recognition is not adopted. Instead, a border is set for the body parts of individual wild animals, and based on the border change rate and the occurrence frequency of the change rate, the abnormal motion behaviors of wild animals are analyzed. Compared with the method of performing image recognition and using related algorithms, it can fully reduce the resource consumption of the analysis work and ensure the analysis accuracy. Description of the Drawings

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the following will briefly introduce the drawings required for use in the embodiments of this application or the prior art. Obviously, the following descriptions are only some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. The drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification, together with the following specific embodiments, to explain the present disclosure, but do not constitute a limitation to the present disclosure. In the drawings: Figure 1 It is a flowchart of an intelligent analysis method for abnormal behaviors of wild animals provided by an embodiment of this application; Figure 2 It is a schematic diagram of an intelligent analysis system for abnormal behaviors of wild animals provided by an embodiment of this application. Detailed Embodiments

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application. In addition, in the embodiments of the present application, "first", "second", etc. are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence.

[0019] In wildlife protection, the analysis of abnormal behaviors occupies an important position. In the currently adopted technical solutions, it is mainly based on the analysis of the tracks of wild animals to determine abnormal behaviors, and this kind of track analysis is usually short-distance tracks. For long-distance track tracking, although it has been used, it is limited by cost and has a relatively narrow scope of use. In addition, in the current analysis, it is mainly based on methods such as manual image analysis or machine vision analysis to analyze abnormal behaviors. The former requires a large amount of human resources, and the latter has low accuracy in many scenarios, such as low-light environments. At the same time, in the existing technical solutions, the image acquisition system adopted is fixedly set, but these fixed devices are very likely to deviate from the migration route of wild animals and cannot achieve the tracking of abnormal behaviors. At the same time, this method has poor results in the analysis of individual abnormal behaviors of wild animals and is difficult to achieve accurate analysis of wild animal abnormal behaviors from the population to the individual. The purpose of this application is to solve various problems existing in the prior art.

[0020] The present application discloses an intelligent analysis method for abnormal behaviors of wild animals, as Figure 1 shown, which is a flowchart of an intelligent analysis method for abnormal behaviors of wild animals provided by an embodiment of the present application, including: S110. Based on a static image acquisition system, identify the types of wild animal populations entering the monitoring area.

[0021] S120. Based on the types of the wild animal populations, obtain the wild animal populations with abnormal tracks.

[0022] S130. Based on the static image acquisition system, obtain multi-directional side views of the wild animal populations with abnormal tracks.

[0023] S140. Based on a dynamic image acquisition system, obtain top views of the wild animal populations with abnormal tracks.

[0024] S150. Based on the top views of the wild animal populations with abnormal tracks, identify the wild animal individuals with abnormal tracks.

[0025] S160. Set a part border for the individual wild animal with abnormal tracks, and obtain the change rate of the part border based on the part border.

[0026] S170. Obtain the abnormal behaviors and causes of wild animals based on the change rate of the part border.

[0027] The purpose of all the above steps is to analyze and determine the abnormal track behaviors of wild animals. After determining the wild animal population with abnormal tracks, determine the abnormal behaviors of individuals within this population. At the same time, in the analysis of abnormal behaviors, instead of using image recognition algorithms, analyze the abnormal behaviors based on the change rate of the border of wild animals, and determine the specific manifestations and causes of abnormal behaviors.

[0028] Next, the specific content of all the above steps will be analyzed. Specifically: As described in step S110, the purpose of this step is to, after setting up a static image acquisition system, considering that not all wild animals entering this monitoring range mean their tracks are abnormal, it is necessary to determine the types of wild animal populations entering the monitoring area, thus laying a foundation for the subsequent identification of wild animal populations with abnormal tracks. Specifically: S111. Obtain the types of wild animal populations in the area, and obtain the normal distribution areas of wild animal populations to obtain the normal monitoring areas.

[0029] The purpose of this step is to analyze the types of wild animal populations in the monitored area. At the same time, obtain the habitats of these animals, determine the normal distribution areas of wild animal populations in the monitoring areas of these regions, so that based on these regions, it can be determined whether the wild animals entering these regions have abnormal track manifestations.

[0030] Among them, for all currently set monitoring areas, they are all set as normal monitoring areas.

[0031] Among them, obtain all the types of wild animal populations in the area, and obtain the habits of these wild animal populations, and then screen the habitats of these wild animals, especially the areas for resting, for use as normal monitoring areas.

[0032] Among them, for the areas where wild animals rest, including flat areas, water resource distribution areas, etc., specifically, they are determined by the habits of the animals.

[0033] Among them, the normal monitoring area refers to the area where the static image acquisition system is set. The so-called static image acquisition system means an image acquisition system where the layout position of the image acquisition device does not change.

[0034] In some embodiments, the image acquisition areas that have already been set on the animal migration routes can also be recognized as normal monitoring areas for obtaining the tracks of wild animals.

[0035] Among them, for image acquisition devices, it includes various devices such as monitoring devices with normal light and infrared light devices.

[0036] S112. Based on the normal distribution area, obtain the abnormal distribution area of the wild animal population.

[0037] The purpose of this step is that on the basis of obtaining the normal monitoring area, it is obviously not sufficient to analyze the tracks of wild animals. Therefore, it is also necessary to analyze the abnormal distribution areas that exist in the daily activities of wild animals, so as to further judge whether the wild animals have reached the abnormal areas.

[0038] Among them, the abnormal distribution area refers to the area that wild animals usually do not reach under normal circumstances during their daily activities, or the corresponding area where the reasons for wild animals to reach this area are usually problems such as wild animal populations getting lost or misidentifying road signs.

[0039] Among them, based on the path formed when wild animals migrate in this area and the continuation of this path, set the abnormal distribution area.

[0040] Among them, for all abnormal distribution areas, they also need to be determined according to the selection criteria of the habitats and rest areas of wild animals.

[0041] Among them, for the abnormal distribution areas of the wild animal population, they are all areas where monitoring devices have not been set yet.

[0042] Among them, for the abnormal distribution area, it needs to be determined based on all wild animal population types in this area, and the abnormal distribution area is screened based on the historical abnormal migration routes of wild animals.

[0043] S113. Based on the distribution spacing between the normal distribution area and the abnormal distribution area, obtain the abnormal degree of the distribution area of the wild animal population.

[0044] The purpose of this step is that during the migration and daily activities of wild animals, the farther the distance between the abnormal distribution area and the normal distribution area they enter, the higher the severity of the abnormal behavior of the population. Therefore, in the analysis of abnormal behavior, it is obviously necessary to analyze the severity of the abnormal behavior corresponding to the abnormal distribution area where it is located. Therefore, corresponding methods need to be used to set the abnormal degree of the distribution area.

[0045] Among them, the distance between the abnormal distribution area and the normal distribution area is obtained. The method for determining the distance is to determine it according to the straight-line distance between the two, and for the abnormal distribution area, it is necessary to calculate with the nearest normal distribution area. The abnormality determination equation is: ; Among them, represents the abnormality of the distribution area, D represents the distance between the abnormal distribution area and its nearest normal distribution area.

[0046] In some embodiments, it is determined according to the available routes of wild animals between the normal distribution area and the abnormal distribution area, and the selection frequency of wild animals for each available route is analyzed for abnormality. The abnormality determination equation at this time is: ; Among them, represents the abnormality of the distribution area, represents the i selection probability of the th available route, i represents the distance of the i th available route, j represents the total number of indexes of the available routes. For this equation, it is necessary to determine all paths between all normal distribution areas and abnormal distribution areas, and determine the selection probability and path distance of all paths. In addition, considering that some wild animals may open up new travel routes during migration or when their tracks are abnormal, it is possible to directly connect the normal distribution area and the abnormal distribution area, and then determine the positions where wild animals can open up new paths between the two connections and calculate the distance.

[0047] Among them, on the basis of obtaining the abnormality, corresponding abnormality numerical values are associated with the abnormal distribution areas.

[0048] S114. Set a static image acquisition system for the abnormal distribution area to obtain an abnormal monitoring area.

[0049] The purpose of this step is that after obtaining the abnormal distribution area, it is obvious that an image acquisition system needs to be set for such areas to monitor wild animals entering the area.

[0050] Among them, at the edge area of the abnormal distribution area, a static image acquisition system is set, and this abnormal distribution area is the abnormal monitoring area.

[0051] Among them, a static image acquisition system is set at the edge of the abnormal distribution area, or when the area of the abnormal distribution area is large, the static image acquisition systems are circularly distributed to achieve image acquisition of the wild animal population in this area.

[0052] Among them, for the static image acquisition system, the devices set include video surveillance devices, photographing devices, etc., and at least include two types, namely, infrared photographing devices and natural light photographing devices.

[0053] S115. Obtain the entry status of wild animals in the normal monitoring area and the abnormal monitoring area, and identify the wild animal population type.

[0054] The purpose of this step is that after setting the normal monitoring area and the abnormal monitoring area, when wild animals appear in the abnormal monitoring area, obviously this wild animal population is a wild animal population with abnormal tracks. However, for the wild animal population that enters the normal monitoring area, the possible situation is that a certain wild animal will not enter this area under normal circumstances. Therefore, it is also necessary to analyze whether the wild animals that enter the normal monitoring area are wild animal populations with abnormal tracks. Therefore, it is necessary to identify the wild animal population type to lay a foundation for the subsequent analysis process.

[0055] Among them, for the wild animals that enter the abnormal monitoring area, it is necessary to identify the wild animal population type based on information such as images and body sizes.

[0056] Among them, for the wild animals that enter the normal monitoring area, it is also necessary to identify the wild animal population type based on information such as images and body sizes.

[0057] As described in step S120, the purpose of this step is that after obtaining the wild animal population type and the monitoring area information, it is necessary to determine the wild animal population with abnormal tracks that can be determined based on these two types of information. Specifically: S121. Set the wild animal population type that normally enters the normal monitoring area to obtain the normal wild animal population type.

[0058] The purpose of this step is to determine whether the wild animal population type that enters the normal monitoring area is reasonable. Obviously, it is also necessary to set a reference parameter for comparison, and this parameter is obviously the wild animal population type that can normally enter this area in the normal monitoring area.

[0059] Among them, based on the habitats and resting areas of the wild animal populations corresponding to the normal monitoring area, obtain the wild animal population type in this area, and establish an association between the wild animal population type information and the normal monitoring area.

[0060] In some embodiments, a mutual conversion system is set up between the abnormal monitoring area and the normal monitoring area. Specifically, when the types of wild animal populations entering a certain abnormal monitoring area exist for a long time and the entry frequency is higher than the set threshold, then this abnormal monitoring area is converted into a normal monitoring area. Similarly, when no wild animal population enters a certain normal monitoring area for a long time, or a certain type of wild animal population has entered for a long time before, but within the monitored time, this wild animal population does not enter for a long time, or the entry frequency is not higher than the threshold, then this normal monitoring area is converted into an abnormal monitoring area.

[0061] Among them, for the types of wild animal populations that normally enter the normal monitoring area, they can be set based on historical data.

[0062] S122. Compare the types of wild animal populations entering the normal monitoring area with the normal wild animal population types. If they are different, then the wild animal population entering the normal monitoring area is a wild animal population with abnormal tracks, and a wild animal population with abnormal tracks in the normal monitoring area is obtained.

[0063] The purpose of this step is that when a normal monitoring area is set, it does not mean that all the wild animal populations entering this area have normal tracks. For example, for two wild animal populations of the same kind, two normal monitoring areas belong to the habitats occupied by the two populations, but when one population enters the habitat of the other population without detecting a fight between the two populations and not during the animal estrus period, then obviously this population has an abnormal track problem. Therefore, this step is to analyze the abnormal track situation of the wild animal populations in the normal monitoring area.

[0064] Among them, for the normal monitoring area, the types of normal wild animal populations entering it are set. Specifically, they are set based on the old monitoring data and population information.

[0065] Among them, after a wild animal population is found to enter the normal monitoring area, based on the analysis of the characteristics of this population, this information is compared with the normal wild animal population types.

[0066] Among them, if it is found that these two types of information cannot correspond, then the newly entered wild animal population is a wild animal population with abnormal tracks.

[0067] S123. Obtain the types of wild animal populations in the abnormal monitoring area and obtain the wild animal populations with abnormal tracks in the abnormal monitoring area.

[0068] The purpose of this step is that, considering that the abnormal monitoring area is a newly built area and the construction basis of such areas is that wild animals usually do not enter, or once they enter, it represents abnormal behavior. Therefore, it is necessary to determine the types of wild animal populations in the abnormal monitoring area to determine the wild animal populations with abnormal tracks.

[0069] Among them, the abnormal monitoring area analyzes whether there are wild animal populations in the current monitoring area. If it is found that they enter, the monitoring system obtains the information of the wild animal populations.

[0070] Among them, for wild animal populations, it does not simply refer to different types of wild animals, but also includes different populations within the same type of wild animal.

[0071] Among them, corresponding settings need to be made for the information of wild animal populations.

[0072] S124. The wild animal populations with abnormal tracks in the normal monitoring area and the wild animal populations with abnormal tracks in the abnormal monitoring area are both set as wild animal populations with abnormal tracks.

[0073] The purpose of this step is that for wild animal populations with abnormal tracks, for all the set wild animal population information, for the convenience of analysis, they are directly set as wild animal populations with abnormal tracks to obtain more convenient explanations and parameter settings.

[0074] Among them, for the wild animal populations with abnormal tracks in the obtained normal monitoring area and abnormal monitoring area, these two types of wild animal populations are directly determined as wild animal populations with abnormal tracks.

[0075] Among them, overall, it is also necessary to analyze whether the wild animal populations entering the normal monitoring area belong to the same type of wild animals as those in the original population but belong to different populations.

[0076] S125. Obtain the types of wild animal populations with abnormal tracks and obtain the severity of the abnormal tracks of the wild animal populations with abnormal tracks.

[0077] The purpose of this step is that when wild animals enter different monitoring areas, in some cases, the degree of abnormal tracks is very small, but in some cases, this indicates that the degree of abnormal tracks is very large, or even indicates that the wild animals have serious disorientation behavior. Therefore, it is necessary to quantify the severity of the abnormal tracks of the wild animals.

[0078] Among them, when obtaining that a wild animal is in an abnormal track state, obtain the type of this wild animal.

[0079] Among them, when a wild animal is in the abnormal monitoring area, obtain the abnormal degree of the distribution area of this area.

[0080] Among them, when the same type of wild animals has different populations, it is necessary to obtain the information of the populations.

[0081] Among them, for the wild animal populations with abnormal trajectories and their locations, an equation for the severity of abnormal trajectories is jointly established. The equation is: ; Among them, represents the severity of abnormal trajectories, represents the abnormality degree of the distribution area, represents the i th wild animal population arriving at the n th abnormal monitoring area, represents the i th wild animal population arriving at the t th normal monitoring area.

[0082] Among them, in the above two equations, can be determined based on the difficulty of wild animals arriving at the abnormal monitoring area, the complexity of the route, etc. This application does not make any limitations.

[0083] Among them, when a wild animal population is in an abnormal monitoring area, there are two situations. One is that the wild animal enters the abnormal monitoring area; the other situation is that the wild animal enters the normal monitoring area, but in the original monitoring data of this wild animal, it indicates that it is impossible to enter this normal monitoring area. For this situation, this normal monitoring area is obviously also an abnormal monitoring area for this wild animal population. At this time, this normal monitoring area is also recognized as an abnormal monitoring area.

[0084] Among them, in the above two equations, In the determination of, the situation is that the wild animals entering the normal monitoring area and the original wild animals in this area belong to the same species but have different populations. For the specific determination of this data, it can be jointly determined based on parameters such as the fighting occurrence frequency between the two populations, the route deviation degree of the population, and the route difficulty of the wild animal population with abnormal trajectories arriving at this normal monitoring area. This application does not make any limitations.

[0085] As described in step S130, for most wild animals, more information about the body shape, posture, and movement can be obtained based on the side view. Therefore, in specific processing, the determination of the side view should be able to take into account that the layout of the image acquisition devices in the image acquisition system is multi-directional, usually circularly arranged, and the image acquisition devices face the interior of the wild animal habitat. Since the orientations of the image acquisition devices are different, their relative positions with respect to the wild animal population are also different. Therefore, it is obvious that multiple side views can be obtained, and these side views are multi-directional side views. Specifically: S131. Based on the static image acquisition system, obtain multi-directional frame images of the wild animal population of the trail.

[0086] The purpose of this step is to better analyze the wild animal population information. Obviously, using a video-based acquisition device can obtain more information and an image basis for image acquisition. Therefore, in this case, frame images need to be obtained to get the processing result.

[0087] Among them, whether it is a normal monitoring area or an abnormal monitoring area, when a wild animal population is detected entering, use all the image acquisition systems set therein to directly obtain frame images of the wild animal population with abnormal trails.

[0088] Among them, since the static image acquisition system includes image acquisition devices with multiple orientations, at this time, keep the orientations of the image acquisition devices unchanged and directly obtain videos of the wild animal population with abnormal trails in different orientations.

[0089] Among them, based on the acquisition of video information, frame images can be directly obtained from the acquired monitoring videos.

[0090] In some embodiments, the obtained frame images are further analyzed to obtain frame images with a relatively small mutual occlusion range among the wild animals in the population, and such frame images are used as target images.

[0091] In some embodiments, for the set image acquisition devices, the wild animal population with abnormal trails can be continuously photographed instead of using video recording to reduce the channel occupancy during the image transmission process, improve the information transmission efficiency, and avoid packet loss.

[0092] Among them, the so-called multi-directional frame images refer to that due to different orientations of multiple image acquisition devices, for the captured videos, different frame images are acquired by different image acquisitions. The frame images obtained by the image acquisition devices with different orientations are multi-directional frame images. S132. Based on the multi-directional frame images and the orientations of all the image acquisition devices in the static image acquisition system, obtain multi-directional side views.

[0093] The purpose of this step is to process the acquired frame image to obtain the corresponding side view, and then lay the foundation for subsequent individual recognition operations based on the obtained side view.

[0094] Among them, the directions of all static image acquisition devices in the static image acquisition system are obtained, the frame images obtained are obtained, and the side views of the wild animal populations in the frame images are obtained.

[0095] Among them, by obtaining the orientation of the image acquisition device and the position of the wild animal population in all image acquisition devices in the frame image, the angle between the current image acquisition device and the wild animal population can be obtained, which can obviously be used to illustrate the relative orientation between the obtained side view and the corresponding image acquisition device.

[0096] The so-called multi-directional side views refer to side views captured by image acquisition devices in different directions.

[0097] Among them, in the acquired multi-directional side views, it must be ensured that the acquisition time of all images is the same, that is, each group of multi-directional side views corresponds to the same time node.

[0098] As described in step S140, the purpose of this step is that although image processing technology has been developed to decompose groups of objects that have occlusion behaviors, on the one hand, the implementation of this technology is relatively complex and consumes a lot of computing resources; on the other hand, the relative positions between individuals of some wild animal populations, and between two or more individuals, do not change for a long time. In terms of imaging, it will be found that the spatial distribution relationship between two or more animals and the image acquisition device is: individual 1-individual 2-image acquisition device, and individual 2 has caused occlusion from the abdomen to the tail of individual 1. At this time, it is impossible to obtain all the image features of individual 1 from the abdomen to the tail based on the image acquisition device, and it is impossible to analyze the abnormal state. In other words, based on the existing image processing technology, it is easy to cause unrecognizable and undetectable parts in the analysis of abnormal behaviors of individual wild animals. In order to solve the above problems, this step lays the foundation for the subsequent individual abnormal behavior determination work by obtaining a top view. Specifically: S141. After the static image acquisition system obtains the wild animal population with abnormal behavior, the dynamic image acquisition system moves to the corresponding monitoring area.

[0099] The purpose of this step is that considering the need to obtain an overhead view of wild animal populations with abnormal tracks, if elevated devices and monitoring equipment are set up in each monitoring area, the effect is poor both in terms of investment cost and the shooting effect obtained. Therefore, setting up a dynamic image acquisition system has a better effect and is more flexible.

[0100] Among them, after a wild animal population with abnormal tracks is found in the static image acquisition system, the dynamic image acquisition system needs to move to this position, and the equipment used can be a drone, a manned detection aircraft, etc.

[0101] Among them, the dynamic image acquisition system should be able to move to the corresponding monitoring area. Obviously, in its specific operation, the received instruction also needs to include the specific position information of the static image acquisition system.

[0102] Among them, the dynamic image acquisition system is in a standby state daily and can only move when the static image acquisition system discovers a wild animal population with abnormal tracks.

[0103] In some embodiments, a wild animal population with abnormal tracks may just pass by a normal monitoring area. In response to this situation, the static image acquisition system predicts the moving position of this population in the next period of time based on the moving direction of the wild animal population with abnormal tracks, and continuously sends guiding instructions to the dynamic image acquisition system, so as to facilitate the dynamic image acquisition system to search for the target.

[0104] In some embodiments, the static image acquisition system also sends the specific target identifier of the wild animal population with abnormal tracks to the dynamic image acquisition system to facilitate the dynamic image acquisition system to confirm the target.

[0105] S142. The dynamic image acquisition system reaches above the wild animal population with abnormal tracks and continuously obtains an overhead view of the wild animal population with abnormal tracks.

[0106] The purpose of this step is to obtain an overhead view after the dynamic image acquisition system reaches above the wild animal population with abnormal tracks.

[0107] Among them, when the dynamic image acquisition system reaches the corresponding monitoring area, it reaches above the wild animal population and obtains image information.

[0108] Among them, in order to obtain more overhead view information, an image monitoring method is adopted to obtain an overhead shooting image.

[0109] In some embodiments, for the situation where a wild animal population passes by a relevant monitoring area, the dynamic image acquisition system needs to continuously track this population and always be above this population.

[0110] Among them, in order to ensure that the dynamic image acquisition system can perform top-view image acquisition all the time, an infrared image acquisition device also needs to be set up.

[0111] As described in step S150, the purpose of this step is to identify each wild animal individual in the wild animal population with abnormal tracks. The theoretical support for this step is that whether wild animals are moving or resting, individuals usually exist independently of each other, rather than being in a superimposed state with each other. At the same time, even when resting in contact with each other, it is easy to distinguish their distribution, which is related to the living habits of a large number of animals. Therefore, by analyzing the top-view image, each individual can be distinguished. Specifically: S151. Based on the top-view image of the wild animal population with abnormal tracks, label the wild animals within the wild animal population with abnormal tracks to obtain individual labeling information.

[0112] The purpose of this step is that during the analysis of the top-view image of the wild animal population, each individual can be obtained. Considering that in the specific analysis of abnormal behaviors, it is necessary to analyze them separately. Therefore, individual labeling information needs to be set to ensure that in the subsequent analysis of the abnormal behaviors of wild animal individuals, based on this individual labeling information, the wild animal individual object for which the abnormal behavior analysis is to be carried out can be determined.

[0113] Among them, after obtaining the top-view image of the wild animal population, identify each individual in the top-view image.

[0114] Among them, for the labeling information of wild animals, key features of each individual can be identified based on feature acquisition technology, and this feature is one of the individual labeling information.

[0115] Among them, the individual feature information includes length, width, color, etc. In some embodiments, it may even include gait analysis, same action frequency, etc., which are not limited in this application.

[0116] Among them, when the wild animal population with abnormal tracks is resting, its individual labeling information can be the position in the top-view image to achieve the purpose of simplifying the processing of individual labeling information.

[0117] S152. Based on the top-view image of the wild animal population with abnormal tracks, associate the individual labeling information with the top-view image of the wild animal individual with abnormal tracks to obtain wild animal individual information with abnormal tracks.

[0118] The purpose of this step is that after obtaining the top-view image of the wild animal population with abnormal tracks, based on the correspondence between all individuals and the corresponding individual labeling information therein, information labels can be set for all individuals.

[0119] Among them, after obtaining the individual annotation information, the corresponding relationships can be directly established for each recognized individual, so as to form the association between the wild animal individuals with abnormal trajectories and the corresponding individual annotation information.

[0120] Among them, in the recognition of all wild animal individuals with abnormal trajectories, the corresponding individual annotation information can be determined based on various information such as the individual orientation, the aspect ratio, length, width, and color of the animal individual, and the two are directly associated.

[0121] In some embodiments, if it is found that all animals in the wild animal population with abnormal trajectories are in a resting state, then numbers are directly set for all of them, and then the individual annotation information association is performed based on the set numbers.

[0122] As described in step S160, the purpose of this step is that the purpose of this application is to determine the abnormal behavior of animal individuals by using the change rate of the border of each part of the wild animal body. Therefore, it is obvious that the border of each part of the wild animal individual with abnormal trajectory needs to be set, and then the abnormal behavior can be analyzed based on the change rate of the part border. Therefore, the purpose of this step is to reasonably set the part border. Specifically: S161. Obtain the body parts of the wild animal individual with abnormal trajectory, and set the side view part border on the multi-directional side view.

[0123] The purpose of this step is that after obtaining the multi-directional side views of wild animals, the corresponding part borders can be set for the animal body parts in each multi-directional side view. At the same time, in this step, multiple different part borders can be set for the same part, and based on this part border, the analysis accuracy of the abnormal state of the same part can be provided.

[0124] Among them, for all multi-directional side views, the unobstructed parts are recognized, then the edge contour line of this part is obtained, and the outermost edge of the contour line is obtained, and the border is set based on all the outermost edges.

[0125] In some embodiments, according to the parts involved in the formation of some actions of wild animals, all the parts that generate such actions are included in the same part border. For example, for even-toed ungulates, obviously the neck and head need to act together when they turn their heads, so at this time, the neck and head can be included in the same part border.

[0126] Among them, for the multi-directional side view, corresponding part borders need to be set for all the unobstructed information in the multi-directional side view, but for the occluded area, no setting is performed, but it is judged based on the subsequent top view border.

[0127] In some embodiments, the position of the side view part border is also determined for subsequent determination of the action amplitude.

[0128] Among them, for the obtained multi-directional side views, in the setting of all part borders therein, it is necessary to set the side view part borders for all other side views based on the part decomposition scheme in one of the side views.

[0129] Among them, for the multi-directional side views, after determining all the part decomposition schemes, the outermost edges of the relevant parts in the multi-directional side views need to be identified, and then the side view part borders are set based on the outermost edges.

[0130] Among them, for the side view part borders, it is also necessary to introduce the individual annotation information and part information of the wild animal individual, and associate these two pieces of information with the lateral part borders.

[0131] S162. Obtain the top view of the wild animal individual with abnormal trajectory, and set the top view part border based on the body part.

[0132] The purpose of this step is that in the analysis of the abnormal behavior of the wild animal individual with abnormal trajectory, on the one hand, if only the side view part border of the wild animal individual is used for the analysis of abnormal behavior, it is difficult to identify the abnormal behavior of some animals in the top view direction, which will lead to defects in the analysis work of abnormal behavior; on the other hand, the side view part border is often incomplete, such as there are occluded areas, then the side view part border cannot be set obviously. In view of this situation, it is necessary to analyze the abnormal behavior based on the part border of the top view. Therefore, it is necessary to set the top view of the wild animal individual with abnormal trajectory.

[0133] Among them, after obtaining the information of the wild animal individual, according to the decomposition rule of the test part border of the wild animal, each part of the animal individual in the top view is decomposed, and the part border is set.

[0134] Among them, in the multi-directional side views of the wild animal individual with abnormal trajectory, if there are occluded parts therein, the occluded parts are identified, and the parts can be divided to set the occluded top view part border.

[0135] Among them, for the top view part border, it is also necessary to set the corresponding part information for the border, and at the same time, it also needs to include the individual annotation information of the wild animal individual with abnormal trajectory, so as to realize the association among the individual information, part information and part border.

[0136] S163. Couple the side-view part border and the top-view part border of the wild animal individual with abnormal traces to obtain a part border.

[0137] The purpose of this step is that for the side-view and top-view part borders of the wild animal individual with abnormal traces obtained, these two types of borders essentially jointly explain the abnormal behaviors of each part in the side-view direction and the top-view direction. Therefore, in order to be able to analyze the abnormal states from these two directions simultaneously, it is necessary to perform coupling processing on the borders among them, so as to jointly achieve the simultaneous analysis of the abnormal states in the side-view direction and the top-view direction.

[0138] Among them, for the obtained side-view part border and top-view part border, it is necessary to analyze the individual annotation information and body part information of the two types of part borders to determine whether the corresponding information of these two part borders exists simultaneously. If it exists, obtain the overlapping sides of the two types of part borders to achieve coupling.

[0139] Among them, if it is found that the individual annotation information and body part information corresponding to the obtained test part border and top-view part border do not match, that is, when at least one of the individual annotation information and body part information corresponding to a certain side-view part border does not exist, then this top-view part border does not need to be further coupled, but is considered to be the result of coupling processing itself.

[0140] Among them, for the part border that has undergone coupling processing, determine it as the part border.

[0141] S164. Obtain all the side change rates of the part border to obtain the part border change rate.

[0142] The purpose of this step is that for all part borders, considering that when an animal makes a movement, the performance of its part border will inevitably change. Therefore, action recognition can be based on the change rate of the border. At the same time, when some abnormal behaviors occur, its action will also be deformed. Based on the change rate, it can be determined whether each action is deformed. If it is found that there is deformation, then individual abnormal behaviors can be recognized.

[0143] Among them, for the case where there are both side-view part borders and top-view part borders and after coupling processing of the two, it is necessary to perform processing on the change rate of one of the part borders, and at the same time, analyze the change rate of the other part border coupled with it to determine whether the change rates of the two part borders correspond. The equation is: ; Among them, 、 、 and respectively represent the length change rate, width change rate of the side view part border, length change rate and width change rate of the top view part border; and respectively represent the length of the side view part border at the action termination time and action start time when the individual of the wild animal with abnormal tracks makes an action; and respectively represent the width of the side view part border at the action termination time and action start time when the individual of the wild animal with abnormal tracks makes an action; and respectively represent the length of the top view part border at the action termination time and action start time when the individual of the wild animal with abnormal tracks makes an action; and respectively represent the width of the top view part border at the action termination time and action start time when the individual of the wild animal with abnormal tracks makes an action; and respectively represent the action termination time point and action start time point when the individual of the wild animal with abnormal tracks makes an action; represents the difference threshold of the change rate of the overlapping side line of the side view part border and the top view part border.

[0144] Among them, for the above equation, if it is found that is greater than then it is necessary to further adjust the length parameter of the overlapping side line of the part border and recalculate the change rate of the part border.

[0145] Among them, the change rates of all multi-directional part side views are also analyzed to obtain the change rate parameters of the part borders in all orientations for obtaining the final result.

[0146] Among them, for the extra top view part border, if it cannot be coupled with the side view part border, the change rate for this top view part border can be directly obtained.

[0147] As described in step S170, the purpose of this step is that after determining the border change rate, the action emission form of the wild animal individual with abnormal tracks can be determined. Then, based on the obtained change rate parameters, the current behavior abnormality form can be directly determined, and further, the cause of this abnormal behavior can be determined based on the behavior abnormality form. Specifically: S171. Obtain the occurrence times of the change rates of all part borders of the wild animal individual with abnormal tracks to obtain the frequency of compulsive behavior.

[0148] The purpose of this step is that when a wild animal shows abnormal movements, it does not necessarily mean that the individual wild animal has abnormal behavior. For example, when an artiodactyl rubs its horns and abdomen, if this movement is considered an abnormal movement during judgment, and it only occurs once in a long time, then this movement is obviously sporadic. However, if this movement occurs frequently, it means that this movement is an abnormal behavior. Therefore, it is necessary to determine the occurrence frequency of relevant behaviors to determine whether an abnormal movement is an abnormal behavior.

[0149] Among them, for the number of occurrences of a specific behavior, it is necessary to determine the occurrence of the change rate of the part border and analyze the occurrence frequency of this change rate. If the number of occurrences is not less than 2 times (obviously, it can also be set to other numbers) and the frequency is high (the actual occurrence frequency can be compared with a preset frequency), the behavior made by the wild animal is set as a compulsive behavior.

[0150] In some embodiments, the abnormal behavior of an individual can also be directly obtained, and the change rate of the part border can be obtained based on this abnormal behavior. When it is found that the change rate of the part border corresponding to a certain action is the same as that of the set abnormal behavior, it is considered that the animal has a compulsive behavior.

[0151] Among them, for the number of occurrences of the change rate of the part border, it can also be determined based on the occurrence frequency of the same action. When the occurrence frequency is high or exceeds the set threshold, it is considered that the action corresponding to the change rate of the part border is a compulsive behavior.

[0152] In some embodiments, it is found that an individual is completely blocked by other individuals, that is: the lateral part border cannot be obtained for this animal. In this case, only the change rate of the top-down part border is analyzed, and the occurrence frequency of compulsive behavior is determined.

[0153] S172. Compare the occurrence frequency of the compulsive behavior with the preset occurrence frequency of the compulsive behavior. If the occurrence frequency of the compulsive behavior is not lower than the preset occurrence frequency of the specific behavior, determine that the wild animal individual with abnormal tracks has abnormal behavior.

[0154] The purpose of this step is to judge whether the wild animal individual with abnormal tracks has abnormal behavior and lay a foundation for subsequent determination of abnormal behavior and cause analysis work.

[0155] Among them, a threshold is set for the compulsive behavior to obtain the preset occurrence frequency of the compulsive behavior, and the measured results are compared to determine whether the wild animal individual with abnormal tracks has abnormal behavior.

[0156] Among them, for the preset occurrence frequency of compulsive behaviors, it can be obtained according to the occurrence frequency of compulsive behaviors corresponding to the obtained abnormal behaviors, that is, it can be determined according to the habits of wild animals.

[0157] S173. Sort the occurrence frequencies of the part border change rates of the wild animal individuals with abnormal behaviors and abnormal traces, and obtain high-frequency abnormal behaviors.

[0158] The purpose of this step is that when a wild animal individual has an abnormal behavior, this abnormal behavior may be set within multiple behaviors. Therefore, it is necessary to determine the high-frequency abnormal behaviors from among multiple action behaviors, so as to obtain the most prominent abnormal behavior characteristics based on this information.

[0159] Among them, sort various occurrence frequencies based on the occurrence frequency of the same part border change rate.

[0160] Among them, based on the set occurrence frequency threshold or the number of occurrence frequencies selected for the occurrence frequency, select high-frequency abnormal behaviors, and the abnormal behaviors corresponding to the obtained occurrence frequencies are high-frequency abnormal behaviors.

[0161] S174. Compare the high-frequency abnormal behaviors, the trace abnormality severity of the wild animal population with abnormal traces, and the causes of the abnormal behaviors of wild animals to obtain the causes of the abnormal behaviors of wild animals.

[0162] The purpose of this step is that for the high-frequency abnormal behaviors that can be obtained and the trace abnormality severity of wild animals, based on these two pieces of information, jointly analyze the causes of the abnormal behaviors of wild animals.

[0163] Among them, set up a retrieval table for the causes of abnormal behaviors, which includes the trace abnormality severity of wild animals, establish the high-frequency abnormal behaviors included under this severity, and for each high-frequency abnormal behavior, set the corresponding cause of the abnormal behavior.

[0164] Among them, for the trace abnormality severity of each wild animal population with abnormal traces, also set the corresponding cause of the trace abnormality of wild animals.

[0165] Among them, after obtaining the trace abnormality severity of the wild animal population with abnormal traces, the cause of the trace abnormality can be directly obtained from this cause retrieval table, and when further obtaining the high-frequency abnormal behaviors, the cause of this specific high-frequency abnormal behavior can be obtained.

[0166] In addition, the present application also discloses an intelligent analysis system for animal abnormal behaviors, such as Figure 2As shown in the figure, it is a schematic diagram of an intelligent animal abnormal behavior analysis system disclosed in an embodiment of the present application. Specifically: An intelligent animal abnormal behavior analysis system includes a static image acquisition system, a dynamic image acquisition system, an animal feature recognition system, a wild animal population recognition system with abnormal trajectories, a wild animal individual recognition system with abnormal trajectories, a wild animal individual part decomposition system with abnormal trajectories, an abnormal behavior analysis system, and a database, and is characterized in that: The static image acquisition system is connected to the animal feature recognition system, and is used to acquire wild animal populations in the normal monitoring area and the abnormal monitoring area, and send the obtained monitoring information to the animal feature recognition system; The animal feature recognition system is connected to the wild animal population recognition system with abnormal trajectories, and is used to acquire wild animal features in the monitoring area; The wild animal population recognition system with abnormal trajectories is also connected to the dynamic image acquisition system, and is used to obtain wild animal population information with abnormal trajectories based on the wild animal features. After obtaining the wild animal population information with abnormal trajectories, it sends a shooting instruction and monitoring information to the image acquisition system; The wild animal individual recognition system with abnormal trajectories is simultaneously connected to the wild animal population recognition system with abnormal trajectories and the dynamic image acquisition system, and is used to acquire images of wild animals with abnormal trajectories; The wild animal individual part decomposition system with abnormal trajectories is connected to the wild animal individual recognition system with abnormal trajectories, and is used to obtain the part borders of wild animals with abnormal trajectories; The abnormal behavior analysis system is connected to the wild animal individual part decomposition system with abnormal trajectories and the wild animal population recognition system with abnormal trajectories, and is used to obtain the types and causes of abnormal behaviors of the wild animal population with abnormal trajectories and the wild animals with abnormal trajectories; The abnormal behavior analysis system is also connected to the database, and the database is used to store the corresponding relationships between the change rates of the part borders of wild animals with abnormal trajectories, the severity of abnormal trajectories of wild animal populations with abnormal trajectories, and the causes of abnormal behaviors of wild animals.

[0167] It also includes: The static image acquisition system includes a normal monitoring area and an abnormal monitoring area. The normal monitoring area is an existing image acquisition area, and the abnormal monitoring area is other image acquisition areas delimited based on the existing image acquisition area; In the normal monitoring area and the abnormal monitoring area, a plurality of image acquisition devices are arranged on the edge lines of the monitoring area.

[0168] The beneficial effects of the present application include: 1. Reduced the implementation cost of analyzing abnormal behaviors of wild animals. In the technical solution of this application, a static and dynamic image acquisition system is established. The static image acquisition system makes full use of the existing image acquisition systems. Meanwhile, based on the existing image acquisition systems, abnormal monitoring areas are selected, and corresponding image acquisition systems are established, achieving effective application of the existing resources. At the same time, the dynamic image acquisition system can reach the monitoring area by itself for image acquisition without fixed installation, enabling the dynamic image acquisition system to be recycled and reused again, reducing the construction cost of the entire system and the implementation cost of the method.

[0169] 2. Broadened the scope of work for analyzing abnormal behaviors of wild animals. In the technical solution of this application, based on the static image acquisition system, wild animal populations with abnormal tracks are obtained. Then, based on the combined application of the static and dynamic image acquisition systems, the abnormal behaviors of individual wild animals with abnormal tracks are analyzed, thus realizing the combined analysis of abnormal behaviors of populations and individuals and broadening the coverage scope of the work for analyzing abnormal behaviors.

[0170] 3. Reduced the resource consumption of the work for analyzing abnormal behaviors of wild animals. In the technical solution of this application, for the process of analyzing abnormal behaviors of wild animals, appearance analysis based on image recognition is not adopted. Instead, a border is set for the body parts of individual wild animals, and based on the border change rate and the occurrence frequency of the change rate, the abnormal motion behaviors of wild animals are analyzed. Compared with the method of performing image recognition and using related algorithms, it can fully reduce the resource consumption of the analysis work and ensure the analysis accuracy.

[0171] Those of ordinary skill in the art can understand that all or part of the steps to implement the above method embodiments can be completed by hardware related to computer program instructions. The aforementioned computer program can be stored in a non-volatile storage medium. When the computer program is executed, it performs the steps including the above method embodiments. Alternatively, if the above integrated units of the present invention are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a non-volatile storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence, or the part that makes a contribution to the prior art can be embodied in the form of a software product. This computer software product is stored in a non-volatile storage medium and includes several instructions to enable an electronic device (which can be a personal computer, a server, a network device, etc.) to execute all or part of the methods described in various embodiments of the present invention.

[0172] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention.

Claims

1. An intelligent analysis method for abnormal behaviors of wild animals, characterized in that, The method includes: Based on a static image acquisition system, identifying the types of wild animal populations entering the monitoring area; Based on the types of wild animal populations, obtaining the wild animal populations with abnormal trajectories; Based on the static image acquisition system, obtaining multi-directional side views of the wild animal populations with abnormal trajectories; Based on a dynamic image acquisition system, obtaining top views of the wild animal populations with abnormal trajectories; Based on the top views of the wild animal populations with abnormal trajectories, identifying individual wild animals with abnormal trajectories; Setting part borders for the individual wild animals with abnormal trajectories, and obtaining the change rate of the part borders based on the part borders; Based on the change rate of the part borders, obtaining the abnormal behaviors and causes of wild animals.

2. The intelligent analysis method for abnormal behaviors of wild animals according to claim 1, wherein The step of, based on a static image acquisition system, identifying the types of wild animal populations entering the monitoring area includes: Obtaining the types of wild animal populations in the area, and obtaining the normal distribution areas of the wild animal populations to obtain normal monitoring areas; Based on the normal distribution areas, obtaining the abnormal distribution areas of the wild animal populations; Based on the distribution spacing between the normal distribution areas and the abnormal distribution areas, obtaining the abnormality degree of the distribution areas of the wild animal populations; Setting a static image acquisition system for the abnormal distribution areas to obtain abnormal monitoring areas; Obtaining the entry status of wild animals in the normal monitoring areas and the abnormal monitoring areas, and identifying the types of wild animal populations.

3. The intelligent analysis method for abnormal behaviors of wild animals according to claim 1, wherein The step of, based on the types of wild animal populations, obtaining the wild animal populations with abnormal trajectories includes: Setting the types of wild animal populations that normally enter for the normal monitoring areas to obtain the types of normal wild animal populations; Comparing the types of wild animal populations entering the normal monitoring areas with the types of normal wild animal populations. If they are different, the wild animal populations entering the normal monitoring areas are the wild animal populations with abnormal trajectories, and obtaining the wild animal populations with abnormal trajectories in the normal monitoring areas; Obtaining the types of wild animal populations in the abnormal monitoring areas, and obtaining the wild animal populations with abnormal trajectories in the abnormal monitoring areas; Both the wild animal populations with abnormal trajectories in the normal monitoring areas and the wild animal populations with abnormal trajectories in the abnormal monitoring areas are set as the wild animal populations with abnormal trajectories; Obtaining the types of wild animal populations with abnormal trajectories, and obtaining the severity of the abnormal trajectories of the wild animal populations with abnormal trajectories.

4. The intelligent analysis method for abnormal behaviors of wild animals according to claim 1, wherein, The step of, based on the static image acquisition system, obtaining multi-directional side views of the wild animal populations with abnormal trajectories includes: Based on the static image acquisition system, obtaining multi-directional frame images of the wild animal populations of the trajectories; Based on the multi-directional frame images and the orientations of all image acquisition devices in the static image acquisition system, obtaining multi-directional side views.

5. The intelligent analysis method for abnormal behaviors of wild animals according to claim 1, wherein The step of, based on a dynamic image acquisition system, obtaining top views of the wild animal populations with abnormal trajectories includes: After the static image acquisition system obtains the wild animal populations with abnormal trajectories, the dynamic image acquisition system moves towards the corresponding monitoring areas; The dynamic image acquisition system reaches above the wild animal populations with abnormal trajectories and continuously obtains top views of the wild animal populations with abnormal trajectories.

6. The intelligent analysis method for abnormal behaviors of wild animals according to claim 1, wherein, The top view of the wild animal population based on the abnormal track, identifying wild animal individuals with abnormal tracks, includes: Based on the top view of the wild animal population with abnormal tracks, labeling the wild animals within the wild animal population with abnormal tracks to obtain individual labeling information; Based on the top view of the wild animal population with abnormal tracks, associating the individual labeling information with the top view of the wild animal individuals with abnormal tracks to obtain wild animal individual information with abnormal tracks.

7. The intelligent analysis method for abnormal behaviors of wild animals according to claim 1, characterized in that, The setting of part borders for the wild animal individuals with abnormal tracks and obtaining the change rate of the part borders based on the part borders includes: Obtaining the body parts of the wild animal individuals with abnormal tracks and setting side view part borders on the multi-directional side view; Obtaining the top view of the wild animal individuals with abnormal tracks and setting top view part borders on the top view based on the body parts; Performing coupling processing on the side view part borders and the top view part borders of the wild animal individuals with abnormal tracks to obtain part borders; Obtaining the change rates of all the side lines of the part borders to obtain the change rate of the part borders.

8. The intelligent analysis method for abnormal behaviors of wild animals according to claim 1, characterized in that The obtaining of the abnormal behaviors and causes of wild animals based on the change rate of the part borders includes: Obtaining the occurrence times of the change rates of all the part borders of the wild animal individuals with abnormal tracks to obtain the occurrence frequency of compulsive behaviors; Comparing the occurrence frequency of the compulsive behaviors with the preset occurrence frequency of compulsive behaviors. If the occurrence frequency of the compulsive behaviors is not lower than the preset occurrence frequency of specific behaviors, determining that the wild animal individuals with abnormal tracks have abnormal behaviors; Sorting the occurrence frequencies of the change rates of the part borders of the wild animal individuals with abnormal behaviors to obtain high-frequency abnormal behaviors; Comparing the high-frequency abnormal behaviors, the severity of the abnormal tracks of the wild animal population with abnormal tracks, and the causes of the abnormal behaviors of wild animals to obtain the causes of the abnormal behaviors of wild animals.

9. A wild animal abnormal behavior intelligent analysis system for performing the wild animal abnormal behavior intelligent analysis method according to any one of claims 1 to 8, including a static image acquisition system, a dynamic image acquisition system, an animal feature recognition system, a wild animal population recognition system with abnormal tracks, a wild animal individual recognition system with abnormal tracks, a wild animal individual part decomposition system with abnormal tracks, an abnormal behavior analysis system, and a database, characterized in that: The static image acquisition system is connected to the animal feature recognition system and is used to acquire wild animal populations in the normal monitoring area and the abnormal monitoring area and send the obtained monitoring information to the animal feature recognition system; The animal feature recognition system is connected to the wild animal population recognition system with abnormal tracks and is used to acquire the features of wild animals in the monitoring area; The wild animal population recognition system with abnormal tracks is further connected to the dynamic image acquisition system and is used to obtain wild animal population information with abnormal tracks based on the wild animal features. After obtaining the wild animal population information with abnormal tracks, it sends a shooting instruction and monitoring information to the image acquisition system; The individual identification system for wild animals with abnormal trajectories is connected to both the population identification system for wild animals with abnormal trajectories and the dynamic image acquisition system, and is used to obtain the individual images of wild animals with abnormal trajectories; The individual part decomposition system for wild animals with abnormal trajectories is connected to the individual identification system for wild animals with abnormal trajectories, and is used to obtain the part boundaries of individual wild animals with abnormal trajectories; The abnormal behavior analysis system is connected to the individual part decomposition system for wild animals with abnormal trajectories and the population identification system for wild animals with abnormal trajectories, and is used to obtain the types and causes of abnormal behaviors of the population of wild animals with abnormal trajectories and individual wild animals with abnormal trajectories; The abnormal behavior analysis system is also connected to the database, which is used to store the corresponding relationships between the change rate of the part boundaries of individual wild animals with abnormal trajectories, the severity of abnormal trajectories of the population of wild animals with abnormal trajectories, and the causes of abnormal behaviors of wild animals.

10. The intelligent analysis system for abnormal behaviors of wild animals according to claim 9, characterized in that, The static image acquisition system further includes: The static image acquisition system includes a normal monitoring area and an abnormal monitoring area. The normal monitoring area is an existing image acquisition area, and the abnormal monitoring area is another image acquisition area delimited based on the existing image acquisition area; In the normal monitoring area and the abnormal monitoring area, a plurality of image acquisition devices are arranged on the edge line of the monitoring area.

Images