A method and system for intelligent analysis of abnormal behavior of wild animals

Abstract

This application discloses an intelligent analysis method and system for abnormal wildlife behavior. The method includes: identifying the type of wildlife population entering a monitored area based on a static image acquisition system; obtaining wildlife populations exhibiting abnormal behavior based on the population type; acquiring multi-directional side views of the wildlife populations exhibiting abnormal behavior based on the static image acquisition system; acquiring top views of the wildlife populations exhibiting abnormal behavior based on a dynamic image acquisition system; identifying individual wildlife exhibiting abnormal behavior based on the top views of the wildlife populations exhibiting abnormal behavior; setting boundary borders for individual wildlife exhibiting abnormal behavior and obtaining the change rate of these boundary borders; and obtaining the abnormal behavior and its causes based on the change rate of these boundary borders. This method solves the problem of insufficient accuracy caused by the failure to simultaneously analyze both behavior patterns and individual behavior in abnormal behavior analysis. It achieves joint analysis of these two types of abnormal behavior and improves the analysis accuracy.

Description

Technical Field

[0001] This application belongs to the field of wildlife behavior monitoring, specifically to a method and system for intelligent analysis of abnormal wildlife behavior. Background Technology

[0002] In wildlife conservation, analyzing abnormal animal behavior is crucial. Currently, image acquisition and processing technologies are being used to analyze these anomalies. However, current methods primarily involve deploying monitoring equipment along wildlife migration routes to track animals and determine if any abnormalities exist. This approach is less effective for analyzing individual wildlife behavior. Furthermore, while wildlife routes may have relatively fixed destinations, their specific patterns can vary, leading to low coupling between the deployed monitoring system and the routes, resulting in poor tracking performance. Moreover, continuously deploying monitoring equipment in wildlife migration areas increases system complexity and construction costs, hindering system robustness and cost control.

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

[0004] To address the problems of excessively high construction costs and insufficient accuracy in analyzing wildlife tracking and individual abnormal behaviors in existing intelligent analysis systems, this application discloses an intelligent analysis method and system for wildlife abnormal behaviors, specifically:

[0005] First aspect:

[0006] A method for intelligent analysis of abnormal behaviors in wild animals, the method comprising:

[0007] Based on a static image acquisition system, identify the types of wild animal populations entering the monitored area;

[0008] Based on the aforementioned wildlife population types, obtain wildlife populations exhibiting unusual behavior patterns;

[0009] Based on the static image acquisition system, a multi-directional side view of the wild animal population with abnormal behavior is obtained;

[0010] A top-down view of the abnormally behaving wild animal population was obtained using a dynamic image acquisition system.

[0011] Based on the top view of the wildlife population with abnormal behavior, identify individual wildlife with abnormal behavior.

[0012] For the wild animal individuals exhibiting abnormal behavior, a boundary border is set for a specific part, and the rate of change of the boundary border is obtained based on the boundary border.

[0013] Based on the change rate of the border of the aforementioned region, the abnormal behavior and causes of wild animals are obtained.

[0014] Optionally, the static image acquisition system for identifying the type of wild animal population entering the monitored area includes:

[0015] Obtain the types of wild animal populations in the region and the normal distribution areas of wild animal populations to obtain the normal monitoring areas;

[0016] Based on the normal distribution area, the abnormal distribution area of ​​the wild animal population is obtained;

[0017] Based on the distribution distance between the normal distribution area and the abnormal distribution area, the distribution area anomaly degree of the wild animal population is obtained;

[0018] A static image acquisition system is set up in the abnormal distribution area to obtain the abnormal monitoring area;

[0019] The system acquires the entry status of wild animals in the normal monitoring area and the abnormal monitoring area, and identifies the wild animal population type.

[0020] Optionally, the step of obtaining wild animal populations exhibiting abnormal behavior based on the wild animal population type includes:

[0021] Set the types of wild animal populations that normally enter the normal monitoring area and obtain the types of normal wild animal populations;

[0022] The types of wild animal populations entering the normal monitoring area are compared with the types of normal wild animal populations. If they are different, the wild animal populations entering the normal monitoring area are wild animal populations with abnormal behavior patterns, and the wild animal populations with abnormal behavior patterns in the normal monitoring area are obtained.

[0023] Obtain the types of wild animal populations within the abnormal monitoring area, and obtain wild animal populations exhibiting abnormal behavior patterns within the abnormal monitoring area;

[0024] Both the abnormal wildlife populations in the normal monitoring area and the abnormal wildlife populations in the abnormal monitoring area are set as abnormal wildlife populations.

[0025] Identify the types of wildlife populations exhibiting abnormal behavior and determine the severity of such abnormal behavior.

[0026] Optionally, acquiring multi-directional side views of the abnormally behaving wild animal population based on the static image acquisition system includes:

[0027] Based on the static image acquisition system, multi-directional frame images of the wildlife population with the tracks are acquired;

[0028] Based on the multi-directional frame image and the orientation of all image acquisition devices in the static image acquisition system, a multi-directional side view is obtained.

[0029] Optionally, acquiring a top-down view of the abnormally behaving wild animal population based on the dynamic image acquisition system includes:

[0030] After the static image acquisition system acquires information about a wild animal population exhibiting abnormal behavior, the dynamic image acquisition system moves to the corresponding monitoring area.

[0031] The dynamic image acquisition system moves above the abnormal wildlife population and continuously acquires a top-down view of the abnormal wildlife population.

[0032] Optionally, identifying individual wild animals exhibiting unusual behavior based on a top-down view of the population includes:

[0033] Based on the top view of the abnormal behavior of the wild animal population, the wild animals in the abnormal behavior population are labeled to obtain individual labeling information;

[0034] Based on the top view of the wild animal population with abnormal behavior, the individual labeling information is associated with the top view of the wild animal individual with abnormal behavior to obtain information about the wild animal individual with abnormal behavior.

[0035] Optionally, setting a boundary border for the individual wild animal exhibiting abnormal behavior and obtaining the rate of change of the boundary border based on the boundary border includes:

[0036] Obtain the body parts of the wild animal individual with abnormal behavior, and set the side view borders on the multi-directional side view;

[0037] Obtain a top view of the individual wild animal exhibiting abnormal behavior, and set a top view border based on the body parts.

[0038] The side view border and the top view border of the wild animal individual with abnormal behavior are coupled to obtain the part border;

[0039] Obtain the rate of change of all edges of the border of the specified region to obtain the rate of change of the region border.

[0040] Optionally, obtaining the abnormal behavior and causes of wild animals based on the change rate of the boundary of the said region includes:

[0041] The frequency of compulsive behavior is obtained by acquiring the number of times the bounding box change rate of all parts of the wild animal individual with abnormal behavior;

[0042] The frequency of the compulsive behavior is compared with the preset frequency of the compulsive behavior. If the frequency of the compulsive behavior is not lower than the preset frequency of the specific behavior, it is determined that the wild animal individual with abnormal behavior has abnormal behavior.

[0043] The frequency of the change rate of the boundary of the part of the wild animal individual with abnormal behavior is sorted to obtain the high-frequency abnormal behavior.

[0044] By comparing the severity of abnormal behavior and the causes of abnormal behavior in the wild animal populations with the frequency of abnormal behavior and the abnormal behavior of the wild animals, the causes of abnormal behavior in wild animals can be obtained.

[0045] The second aspect:

[0046] An intelligent analysis system for abnormal animal behavior, 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 identification system for abnormal behavior patterns, a wild animal individual identification system for abnormal behavior patterns, a wild animal individual part decomposition system for abnormal behavior patterns, an abnormal behavior analysis system, and a database, characterized in that:

[0047] The static image acquisition system and the animal feature recognition system are connected to acquire information on wild animal populations in normal and abnormal monitoring areas, and send the acquired monitoring information to the animal feature recognition system.

[0048] The animal feature recognition system is connected to the abnormal wildlife population recognition system to obtain wildlife features within the monitoring area;

[0049] The abnormal behavior wildlife population identification system is also connected to the dynamic image acquisition system, and is used to obtain abnormal behavior wildlife population information based on the wildlife characteristics. After obtaining the abnormal behavior wildlife population information, the system sends shooting instructions and monitoring information to the image acquisition system.

[0050] The abnormal behavior wildlife individual identification system is connected to both the abnormal behavior wildlife population identification system and the dynamic image acquisition system to acquire images of the abnormal behavior wildlife individuals.

[0051] The system for decomposing the parts of wild animals with abnormal behavior patterns and the system for identifying wild animals with abnormal behavior patterns are connected to obtain the bounding boxes of the parts of wild animals with abnormal behavior patterns.

[0052] The abnormal behavior analysis system is connected to the abnormal behavior individual part decomposition system and the abnormal behavior population identification system for wild animals, and is used to obtain the types and causes of abnormal behaviors of the abnormal behavior population and the abnormal behavior of the abnormal behavior individuals of wild animals.

[0053] The abnormal behavior analysis system is also connected to the database, which stores the rate of change of the body part borders of the wild animal individuals with abnormal behavior, the severity of the abnormal behavior of the wild animal population with abnormal behavior, and the correspondence between the causes of the abnormal behavior of wild animals.

[0054] Optionally, the still image acquisition system further includes:

[0055] The static image acquisition system includes a normal monitoring area and an abnormal monitoring area. The normal monitoring area is the existing image acquisition area, and the abnormal monitoring area is other image acquisition areas defined based on the existing image acquisition area.

[0056] In both the normal monitoring area and the abnormal monitoring area, multiple image acquisition devices are set along the edge line of the monitoring area.

[0057] The beneficial effects of this application include:

[0058] 1. Reduced implementation costs for analyzing abnormal wildlife behavior. The technical solution of this application establishes static and dynamic image acquisition systems. The static image acquisition system fully utilizes existing image acquisition systems and selects abnormal monitoring areas based on these systems, establishing corresponding image acquisition systems to effectively utilize existing resources. Simultaneously, the dynamic image acquisition system can autonomously reach the monitoring area to acquire images without fixed setup, allowing for the recycling and reuse of the dynamic image acquisition system, thus reducing the overall system construction and implementation costs.

[0059] 2. It broadens the scope of analysis of abnormal behavior in wild animals. The technical solution of this application uses a static image acquisition system to obtain data on wild animal populations exhibiting abnormal behavior. Then, based on the combined application of static and dynamic image acquisition systems, the abnormal behavior of individual wild animals exhibiting abnormal behavior is analyzed. This achieves joint analysis of abnormal behavior in both populations and individuals, thus broadening the coverage of abnormal behavior analysis.

[0060] 3. Reduced resource consumption in analyzing abnormal behavior of wild animals. The technical solution of this application does not employ image recognition-based appearance analysis for analyzing abnormal behavior of wild animals. Instead, it sets bounding boxes around individual wild animals and analyzes abnormal behavior based on the rate of change of these bounding boxes and the frequency of their occurrence. Compared to image recognition and related algorithms, this significantly reduces resource consumption while ensuring analytical accuracy. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the embodiments of this application or the prior art will be briefly introduced below. Obviously, the following description is only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. The drawings are used to provide a further understanding of this disclosure and constitute a part of the specification. They are used together with the following detailed description to explain this disclosure, but do not constitute a limitation of this disclosure. In the drawings:

[0062] Figure 1 A flowchart of an intelligent analysis method for abnormal behavior of wild animals provided in this application embodiment;

[0063] Figure 2 This is a schematic diagram of an intelligent analysis system for abnormal wildlife behavior provided in an embodiment of this application. Detailed Implementation

[0064] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Furthermore, in the embodiments of this application, "first," "second," etc., are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0065] In wildlife conservation, the analysis of abnormal behavior plays a crucial role. Current technological solutions primarily rely on analyzing wildlife tracks to identify anomalous behaviors. This track analysis is typically limited to short distances; while long-distance tracking has been used, its application is limited by cost. Furthermore, current analyses mainly rely on manual image analysis or machine vision analysis. The former requires significant human resources, while the latter suffers from low accuracy in many scenarios, such as low-light environments. Additionally, existing image acquisition systems are fixed, which may deviate from wildlife migration routes, hindering the tracking of anomalous behaviors. Moreover, this method is ineffective for analyzing individual wildlife behavior, making it difficult to accurately analyze anomalous behaviors from the population level to the individual. The purpose of this application is to address the various problems existing in the current technologies.

[0066] This application discloses an intelligent analysis method for abnormal behaviors of wild animals, such as... Figure 1 The diagram shown is a flowchart of an intelligent analysis method for abnormal wildlife behavior provided in an embodiment of this application, including:

[0067] S110. Based on a static image acquisition system, identify the types of wild animal populations entering the monitored area.

[0068] S120. Based on the wild animal population type, obtain wild animal populations with abnormal behavior patterns.

[0069] S130. Based on the static image acquisition system, obtain multi-directional side views of the wild animal population with abnormal behavior.

[0070] S140. Based on the dynamic image acquisition system, obtain a top view of the wild animal population with abnormal behavior.

[0071] S150. Based on the top view of the abnormal behavior of the wild animal population, identify individual wild animals with abnormal behavior.

[0072] S160. Set a part border for the wild animal individual with abnormal behavior, and obtain the part border change rate based on the part border.

[0073] S170. Based on the change rate of the border of the said part, obtain the abnormal behavior and cause of the wild animal.

[0074] The purpose of all the above steps is to analyze and identify abnormal behaviors of wild animals, and after identifying the population of wild animals with abnormal behaviors, to identify the abnormal behaviors of individuals within that population. In addition, image recognition algorithms are no longer used in the abnormal behavior analysis; instead, the analysis is based on the change rate of the wild animals' bounding boxes to determine the specific manifestations and causes of the abnormal behaviors.

[0075] The following will analyze the specific details of all the above steps:

[0076] As described in step S110, the purpose of this step is, after the static image acquisition system has been set up, to determine the species of wild animals entering the monitoring area, considering that not all wild animals entering the monitoring area indicate abnormal behavior. This lays the foundation for subsequent identification of wild animal populations exhibiting abnormal behavior. Specifically:

[0077] S111. Obtain the types of wild animal populations in the region and the normal distribution areas of wild animal populations to obtain the normal monitoring areas.

[0078] The purpose of this step is to analyze the types of wildlife populations in the monitored area, obtain the habitats of these animals, determine the normal distribution areas of wildlife populations within the monitored areas, and thus determine whether wildlife entering these areas exhibits abnormal behavior based on these areas.

[0079] All currently configured monitoring areas will be set to normal monitoring areas.

[0080] This involves acquiring all wildlife population types within the region and understanding their habits, then selecting habitats for these wildlife species, especially resting areas, for regular monitoring.

[0081] Among them, the areas used by wild animals for resting include flat areas and areas with abundant water resources, and the specifics are determined by the animals' habits.

[0082] The normal monitoring area refers to the area where the static image acquisition system is set up. The static image acquisition system is an image acquisition system in which the location of the image acquisition equipment does not change.

[0083] In some embodiments, image acquisition areas already set up along animal migration routes can also be considered as normal monitoring areas to obtain the movement tracks of wild animals.

[0084] Image acquisition equipment includes various types of equipment such as monitoring equipment under normal lighting and infrared lighting equipment.

[0085] S112. Based on the normal distribution area, obtain the abnormal distribution area of ​​the wild animal population.

[0086] The purpose of this step is that obtaining the normal monitoring area is obviously not enough to analyze the tracks of wild animals. Therefore, it is also necessary to analyze the abnormal distribution areas of wild animals in their daily activities, so as to further determine whether wild animals have reached abnormal areas.

[0087] Among them, abnormal distribution areas refer to areas that wild animals would not normally reach under normal circumstances, or areas where wild animals usually reach the area due to problems such as disorientation or misidentification of landmarks.

[0088] Among them, abnormal distribution areas are set based on the paths formed by the migration of wild animals in the region and the continuation of these paths.

[0089] Among these, all areas of abnormal distribution must be determined based on the selection criteria for wildlife habitats and rest areas.

[0090] Among them, the areas with abnormal distribution of wild animal populations are all areas where monitoring equipment has not yet been installed.

[0091] For areas of abnormal distribution, it is necessary to determine the distribution based on all wildlife population types within the area and to screen for abnormal distribution areas based on the historical abnormal migration routes of wildlife.

[0092] S113. Based on the distribution distance between the normal distribution area and the abnormal distribution area, obtain the distribution area anomaly degree of the wild animal population.

[0093] The purpose of this step is that, in the migration and daily activities of wild animals, the greater the distance between the abnormal distribution area they enter and the normal distribution area, 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 abnormal behavior corresponding to the abnormal distribution area. Thus, it is necessary to use appropriate methods to set the abnormality degree of the distribution area.

[0094] The method for determining the distance between abnormal and normal distribution regions is based on their straight-line distance, and the abnormal distribution region needs to be compared with the nearest normal distribution region. The anomaly determination equation is:

[0095] ;

[0096] in, Indicates the degree of anomaly in the distribution area. D It represents the distance between the abnormal distribution area and its nearest normal distribution area.

[0097] In some embodiments, the anomaly is determined based on the available wildlife routes between normal and abnormal distribution areas, and an anomaly analysis is performed on the frequency of wildlife selection along each available route. The anomaly determination equation in this case is:

[0098] ;

[0099] in, Indicates the degree of anomaly in the distribution area. Indicates the first i The probability of choosing one of the available routes. Indicates the first i The distance of the available routes, i Indicates the index of available routes. j This represents the total number of available route indices. For this equation, it is necessary to determine all paths between all normal and abnormal distribution areas, and to determine the selection probability and path length for each path. Furthermore, considering that some wild animals may establish new routes during migration or when their behavior is unusual, a method can be used to directly connect normal and abnormal distribution areas, then determine the locations between these two lines where wild animals can establish new routes, and calculate the distances.

[0100] In this process, based on the anomaly degree, each anomaly distribution area is associated with a corresponding anomaly degree value.

[0101] S114. Set up a static image acquisition system for the abnormal distribution area to obtain the abnormal monitoring area.

[0102] The purpose of this step is that, after identifying areas of abnormal distribution, it is obviously necessary to set up image acquisition systems in these areas to monitor wild animals entering them.

[0103] In particular, a static image acquisition system is set up in the edge area of ​​the abnormal distribution area, and this abnormal distribution area is the abnormal monitoring area.

[0104] Among them, static image acquisition systems are set up at the edges of abnormal distribution areas, or in a circular pattern when the area of ​​abnormal distribution is large, to achieve image acquisition of wild animal populations in the area.

[0105] For static image acquisition systems, the devices included include video surveillance equipment, photography equipment, etc., with at least two types: infrared imaging equipment and natural light imaging equipment.

[0106] 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.

[0107] The purpose of this step is to determine whether wild animals entering the normal monitoring area (where they are not normally observed) constitute an anomalous population, given the establishment of normal and abnormal monitoring zones. This requires further analysis to determine if these animals are indeed anomalous. Therefore, identifying the type of wild animal population is crucial for subsequent analysis.

[0108] Among these measures, for wild animals entering abnormally monitored areas, it is necessary to identify the species type of wild animals based on information such as images and body size.

[0109] In addition, for wild animals that enter the normal monitoring area, it is also necessary to identify the type of wild animal population based on information such as images and body size.

[0110] As described in step S120, the purpose of this step is to identify, after obtaining information on the wildlife population type and the monitored area, the wildlife populations exhibiting abnormal behavior that can be determined based on these two types of information. Specifically:

[0111] S121. Set the types of wild animal populations that can enter the normal monitoring area and obtain the types of normal wild animal populations.

[0112] The purpose of this step is to determine whether the types of wildlife populations entering the normal monitoring area are reasonable. Obviously, it is also necessary to set a benchmark parameter for comparison, which is obviously the types of wildlife populations that can normally enter the normal monitoring area.

[0113] Among them, based on the habitat and resting areas of the wild animal populations corresponding to the normal monitoring area, the wild animal population types in the area are obtained, and the wild animal population type information is linked to the normal monitoring area.

[0114] In some embodiments, a conversion system is set up between abnormal monitoring areas and normal monitoring areas. Specifically, when a type of wild animal population enters an abnormal monitoring area for a long period of time and the frequency of entry is higher than a set threshold, the abnormal monitoring area is converted into a normal monitoring area. Similarly, when a normal monitoring area has not seen any wild animal population for a long period of time, or when a certain type of wild animal population has been entering for a long time before, but has not entered for a long period of time during the monitoring period, or the frequency of entry is not higher than the threshold, the normal monitoring area is converted into an abnormal monitoring area.

[0115] Among them, the types of wild animal populations that normally enter the normal monitoring area can be set based on historical data.

[0116] S122. Compare the type of wild animal population entering the normal monitoring area with the type of normal wild animal population. If they are different, the wild animal population entering the normal monitoring area is a wild animal population with abnormal behavior, and the wild animal population with abnormal behavior in the normal monitoring area is obtained.

[0117] The purpose of this step is to address the fact that establishing a normal monitoring area does not guarantee that all wildlife populations entering that area are behaving normally. For example, if two similar wildlife populations occupy different monitoring areas, but one population enters the other's habitat when no fighting is detected between the two populations and when the populations are not in mating season, then this population is clearly exhibiting abnormal behavior. Therefore, this step aims to analyze abnormal behavior patterns among wildlife populations within the normal monitoring area.

[0118] For normal monitoring areas, the types of normal wild animal populations that can enter are set, specifically based on existing monitoring data and population information.

[0119] In this case, once a wild animal population is discovered within a normal monitoring area, the information is compared with the normal wild animal population type based on the analysis of the population's characteristics.

[0120] If these two types of information do not match, the newly entered wildlife population is considered to be a wildlife population with abnormal behavior.

[0121] S123. Obtain the types of wild animal populations within the abnormal monitoring area, and obtain wild animal populations with abnormal behavior patterns within the abnormal monitoring area.

[0122] The purpose of this step is to identify the species of wild animals in the abnormal monitoring area, given that the abnormal monitoring area is a newly established area and is built on the basis of areas that wild animals usually do not enter, or where their entry indicates abnormal behavior.

[0123] Among them, the abnormal monitoring area analysis checks whether there are wild animal populations in the current monitoring area. If an entry is detected, the monitoring system obtains information about the wild animal population.

[0124] In this context, "wildlife populations" does not simply refer to different types of wild animals, but also includes different populations within the same species of wild animal.

[0125] Among these, appropriate settings need to be made for information on wildlife populations.

[0126] S124. The abnormal wildlife populations in the normal monitoring area and the abnormal wildlife populations in the abnormal monitoring area are both set as abnormal wildlife populations.

[0127] The purpose of this step is to directly set all wild animal population information as abnormal behavior groups for easier analysis, so as to obtain more convenient explanations and parameter settings.

[0128] Among them, for the wild animal populations with abnormal behavior in both the normal and abnormal monitoring areas, both types of wild animal populations were directly identified as wild animal populations with abnormal behavior.

[0129] In addition, from an overall perspective, it is also necessary to analyze whether the wild animal populations entering the normal monitoring area belong to the same category of wild animals as the original population, but belong to different populations.

[0130] S125. Obtain the type of wild animal population with abnormal behavior and obtain the severity of the abnormal behavior of the wild animal population with abnormal behavior.

[0131] The purpose of this step is to identify areas where wild animals enter different monitoring zones. Some animals exhibit only minor abnormal behavior, while others show significant abnormality, even indicating severe disorientation. Therefore, it is necessary to quantify the severity of abnormal behavior in wild animals.

[0132] Among them, when a wild animal is in an abnormal behavior state, the type of wild animal is obtained.

[0133] Among them, when wild animals are in an abnormal monitoring area, the anomaly degree of the distribution area in that area is obtained.

[0134] When the same type of wild animal exists, but the populations are different, it is necessary to obtain information about the populations.

[0135] Among them, a severity equation for the abnormal behavior of wild animal populations and their locations is jointly established, which is as follows:

[0136] ;

[0137] in, Indicates the severity of abnormal behavior. Indicates the degree of anomaly in the distribution area. Indicates the first i The population of wild animals reached the n The probability of an abnormal monitoring area. Indicates the first i The population of wild animals reached the t The probability of a normal monitoring area.

[0138] Among the two equations above, The determination of the area can be based on factors such as the difficulty and route complexity of wild animals reaching the abnormal monitoring area, and this application does not impose any restrictions.

[0139] 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 is that the wild animal enters the normal monitoring area, but the original monitoring data indicates that the wild animal could not have entered the normal monitoring area. In this case, the normal monitoring area is obviously also an abnormal monitoring area relative to the wild animal population, and the normal monitoring area is also identified as an abnormal monitoring area.

[0140] Among the two equations above, In determining the situation, the wild animals entering the normal monitoring area belong to the same species as the original wild animals in the area, but the populations are different. The specific determination of this data can be based on parameters such as the frequency of combat between the two populations, the degree of deviation of the populations' routes, and the difficulty of the routes taken by wild animal populations with abnormal behavior to reach the normal monitoring area. This application does not impose any restrictions on these parameters.

[0141] As described in step S130, the purpose of this step is that, for most wild animals, more information about their shape, posture, and movement can be obtained based on side views. Therefore, in the specific processing, it is necessary to be able to determine the side views. Considering that the image acquisition devices in the image acquisition system are arranged in a multi-directional manner, usually in a circular pattern, with the image acquisition devices facing into the wild animal habitat, the relative position of the image acquisition devices to the wild animal population varies depending on their orientation. Therefore, it is obvious that multiple side views can be obtained; these side views are called multi-directional side views. Specifically:

[0142] S131. Based on the static image acquisition system, acquire multi-frame images of the wildlife population with the tracks.

[0143] The purpose of this step is to enable better analysis of wildlife population information. Using video acquisition devices can obviously obtain more information and an image basis for image acquisition. Therefore, in this case, it is necessary to acquire frame images to obtain the processing results.

[0144] In both normal and abnormal monitoring areas, when an intruding wild animal population is detected, the frame images of the wild animal population exhibiting abnormal behavior are directly obtained from all the image acquisition systems set up within the area.

[0145] In this case, since the static image acquisition system contains image acquisition devices facing multiple directions, the orientation of the image acquisition devices is kept unchanged, and videos of wild animal populations with abnormal behavior are directly acquired from different orientations.

[0146] Among them, based on the acquisition of video information, frame images can be directly obtained from the acquired surveillance video.

[0147] In some embodiments, the acquired frame images are further analyzed to obtain frame images in which the wild animals in the population have a small occlusion range with each other, and these frame images are used as target images.

[0148] In some embodiments, for the image acquisition device, continuous image capture can be used to capture images of wild animal populations exhibiting unusual behavior, instead of video recording, in order to reduce channel occupancy during image transmission, improve information transmission efficiency, and avoid packet loss.

[0149] In this context, "multi-directional frame image" refers to a situation where multiple image acquisition devices, facing different directions, capture different frames of a video. The frames obtained from these different oriented image acquisition devices constitute a multi-directional frame image.

[0150] S132. Based on the multi-directional frame image and the orientation of all image acquisition devices in the static image acquisition system, obtain a multi-directional side view.

[0151] The purpose of this step is to process the acquired frame images to obtain the corresponding side view, which can then be used as the basis for subsequent individual identification operations.

[0152] In this process, the orientation of all static image acquisition devices in the static image acquisition system is determined, their captured frame images are obtained, and side views of the wildlife populations within the frame images are acquired.

[0153] Among them, obtaining the orientation of the image acquisition device and the position of the wild animal population in the frame image of all image acquisition devices can obtain the angle between the current image acquisition device and the wild animal population, which can obviously be used to explain the relative orientation between the obtained side view and the corresponding image acquisition device.

[0154] The so-called multi-directional side view refers to the side view captured by the image acquisition device facing different directions.

[0155] In the acquisition of multi-directional side views, it is essential to ensure that all images are acquired at the same time, meaning that each set of multi-directional side views corresponds to the same time node.

[0156] As described in step S140, the purpose of this step is that although image processing techniques for decomposing group objects exhibiting mutual occlusion behavior have been developed, these techniques are complex to implement and consume significant computational resources. Furthermore, in some wild animal populations, the relative positions of individuals, or even multiple individuals, remain unchanged for extended periods. This leads to the following spatial distribution relationship between two or more animals and the image acquisition device: Individual 1 – Individual 2 – Image Acquisition Device. Individual 2 occludes Individual 1 from its abdomen to its tail. In this case, the image acquisition device cannot acquire all image features of Individual 1 from its abdomen to its tail, making it impossible to analyze abnormal states. In other words, existing image processing techniques easily result in unidentifiable and undetectable areas during the analysis of abnormal behavior in wild animals. To address these issues, this step obtains a top-down view, laying the foundation for subsequent individual abnormal behavior determination. Specifically:

[0157] S141. After the static image acquisition system acquires information about a wild animal population exhibiting abnormal behavior, the dynamic image acquisition system moves to the corresponding monitoring area.

[0158] The purpose of this step is to address the need for aerial views when acquiring information about wildlife populations exhibiting unusual behavior. Setting up elevated structures and monitoring equipment in each monitoring area is less effective in terms of both cost and image quality. Therefore, a dynamic image acquisition system is more efficient and flexible.

[0159] Once a population of wild animals exhibiting unusual behavior is detected within the static image acquisition system, the dynamic image acquisition system needs to move to that location. The equipment used can include drones, manned reconnaissance aircraft, etc.

[0160] The dynamic image acquisition system needs to be able to move to the corresponding monitoring area. Obviously, in its actual operation, the received instructions also need to include the specific location information of the static image acquisition system.

[0161] The dynamic image acquisition system is normally on standby and can only be moved when the static image acquisition system detects a wild animal population behaving abnormally.

[0162] In some embodiments, a wild animal population exhibiting unusual behavior may simply be passing through a normal monitoring area. In such cases, a static image acquisition system can predict the population's location over a period of time based on its direction of movement and continuously send guidance instructions to a dynamic image acquisition system, thereby facilitating the dynamic image acquisition system's search for the target.

[0163] In some embodiments, the static image acquisition system also sends specific target identifiers of wildlife populations exhibiting unusual behavior to the dynamic image acquisition system, so that the dynamic image acquisition system can identify the target.

[0164] S142. The dynamic image acquisition system reaches above the abnormally behaving wild animal population and continuously acquires a top-down view of the abnormally behaving wild animal population.

[0165] The purpose of this step is to obtain a top-down view of a wildlife population exhibiting unusual behavior using a dynamic image acquisition system.

[0166] The dynamic image acquisition system, upon reaching the corresponding monitoring area, moves above the wildlife population and acquires image information.

[0167] In order to obtain more information from the top view, video surveillance was used to acquire images taken from above.

[0168] In some embodiments, when a wildlife population is passing through a relevant monitoring area, the dynamic image acquisition system needs to continuously track the population and always remain above it.

[0169] In order to ensure that the dynamic image acquisition system can acquire top-down views at all times, an infrared image acquisition device also needs to be set up.

[0170] As described in step S150, the purpose of this step is to identify each individual wild animal in a population exhibiting abnormal behavior. The theoretical basis for this step is that wild animals, whether moving or resting, typically exist independently rather than in a superimposed state. Furthermore, even when resting and in contact with each other, their distribution is easily distinguishable, which is related to the lifestyle habits of many animals. Therefore, by analyzing the top-down view, each individual can be differentiated. Specifically:

[0171] S151. Based on the top view of the abnormal behavior wild animal population, the wild animals in the abnormal behavior wild animal population are labeled to obtain individual labeling information.

[0172] The purpose of this step is to obtain individual individuals in the top-down view analysis of a wildlife population. Considering that individual analysis is required in specific abnormal behavior analysis, it is necessary to set individual labeling information to ensure that the individual wildlife being analyzed can be identified based on this labeling information in the subsequent abnormal behavior analysis process.

[0173] After obtaining a top-down view of the wildlife population, the individual animals in the view are identified.

[0174] Among them, the labeling information for wild animals can be based on feature acquisition technology to identify the key features of each individual, which is one of the individual labeling information.

[0175] The individual characteristic information includes length, width, color, etc., and in some embodiments, it may even include gait analysis, frequency of the same movement, etc., which are not limited in this application.

[0176] Among them, when wild animal populations exhibiting unusual behavior are at rest, their individual labeling information can be their location in the top-down view, thus simplifying the processing of individual labeling information.

[0177] S152. Based on the top view of the wild animal population with abnormal behavior, associate the individual labeling information with the top view of the wild animal individual with abnormal behavior to obtain information on the wild animal individual with abnormal behavior.

[0178] The purpose of this step is to, after obtaining an overhead view of a wild animal population exhibiting unusual behavior, assign corresponding information tags to each individual based on the mapping between all individuals and their corresponding labeling information.

[0179] After obtaining individual annotation information, the corresponding relationships between each identified individual can be directly established, thereby linking wild animal individuals with abnormal behavior patterns with their corresponding individual annotation information.

[0180] In identifying all wild animal individuals with unusual behavior patterns, the corresponding individual labeling information can be determined based on various factors such as the individual's orientation, aspect ratio, length, width, and color, and these two factors can be directly correlated.

[0181] In some embodiments, if all animals in a wild animal population exhibiting abnormal behavior are found to be at rest, then all individuals can be assigned a number, and then individual labeling information can be associated based on the assigned number.

[0182] As described in step S160, the purpose of this step is to determine abnormal behavior in an individual animal by using the rate of change of the bounding boxes of various parts of the animal's body. Therefore, it is obvious that it is necessary to set bounding boxes for each part of the animal exhibiting abnormal behavior before the analysis of abnormal behavior can be performed based on the rate of change of the bounding boxes. Therefore, the purpose of this step is to reasonably set the bounding boxes. Specifically:

[0183] S161. Obtain the body parts of the wild animal individual with abnormal behavior and set the side view border on the multi-directional side view.

[0184] The purpose of this step is to set corresponding part borders for each part of the animal's body in the multi-view side view of the obtained wildlife. At the same time, multiple different part borders can be set for the same part in this step, and the analysis accuracy of abnormal conditions of the same part can be improved based on the part border.

[0185] In this process, unobstructed parts are identified for all multi-directional side views. Then, the edge contour of the part is obtained, and the outermost edge of the contour is obtained. A border is set based on all the outermost edges.

[0186] In some embodiments, based on the parts involved in the formation of certain movements in wild animals, all parts that generate such movements are included in the same part border. For example, for even-toed ungulates, the neck and head obviously need to work together when they turn their heads, so the neck and head can be included in the same part border.

[0187] Specifically, for multi-directional side views, it is necessary to set corresponding part borders for all unobstructed information in the multi-directional side view, but for obstructed areas, no borders are set, and the judgment is based on the borders of the subsequent top view.

[0188] In some embodiments, the position of the side view border is also determined for subsequent motion range determination.

[0189] In the process of setting the borders of all parts of the obtained multi-directional side view, it is necessary to set the borders of the side view parts based on the part decomposition scheme of one of the side views, and set the borders of the side view parts of all other side views based on the same decomposition scheme.

[0190] For multi-directional side views, after determining all the part decomposition schemes, the outermost edge of the relevant part in the multi-directional side view must be identified, and then the side view part border is set based on the outermost edge.

[0191] For the side view border, it is also necessary to introduce the individual labeling information and part information of the wild animal, and associate these two pieces of information with the side view border.

[0192] S162. Obtain a top view of the individual wild animal with abnormal behavior, and set a top view border based on the body parts.

[0193] The purpose of this step is to address the following issues in the analysis of abnormal behavior in wild animal individuals exhibiting unusual behavior patterns: Firstly, if the analysis is based solely on the side view borders of the wild animal individual, it will be difficult to identify abnormal behaviors in the top view, leading to defects in the analysis. Secondly, the side view borders are often incomplete, such as having occluded areas, making it impossible to set side view borders. In such cases, it is necessary to analyze abnormal behavior based on the borders of the top view. Therefore, it is essential to set a top view for wild animal individuals exhibiting unusual behavior patterns.

[0194] After obtaining information about individual wild animals, the process involves decomposing each part of the animal in the top view according to the decomposition rules for the test part borders of wild animals, and setting the part borders.

[0195] In particular, in the multi-directional side view of wild animal individuals with abnormal behavior, if there are occluded parts, the occluded parts can be identified and divided into parts to set the border of the occluded top view parts.

[0196] Specifically, for the border of the top view area, it is also necessary to set the corresponding area information for the border, and at the same time, it is also necessary to include the individual labeling information of wild animals with abnormal behavior, so as to realize the association between individual information, area information and area border.

[0197] S163. The side view border and the top view border of the wild animal individual with abnormal behavior are coupled to obtain the part border.

[0198] The purpose of this step is to obtain the side and top view bounding boxes of the obtained wild animal individuals with abnormal behavior. These two types of bounding boxes are essentially a joint description of the abnormal behavior of each part in the side and top view directions. Therefore, in order to analyze the abnormal state from both directions at the same time, it is necessary to couple the bounding boxes to achieve the simultaneous analysis of the abnormal state in the side and top view directions.

[0199] Specifically, for the obtained side view and top view bounding boxes, it is necessary to analyze the individual annotation information and body part information of the two types of bounding boxes to determine whether the corresponding information of the two bounding boxes exists simultaneously. If so, the overlapping edges of the two types of bounding boxes are obtained to achieve coupling.

[0200] If it is found that the individual annotation information and body part information corresponding to the obtained test part border and top view border do not correspond, that is, if at least one of the individual annotation information and body part information corresponding to a certain side view border is missing, then the top view border does not need to be further coupled, but is considered to be the result of coupling processing itself.

[0201] Among them, the part borders that have undergone coupling processing are determined as part borders.

[0202] S164. Obtain the change rate of all edges of the border of the part, and get the change rate of the border of the part.

[0203] The purpose of this step is to recognize that the appearance of the bounding boxes of all parts will inevitably change when the animal moves. Therefore, action recognition can be performed based on the rate of change of the bounding boxes. At the same time, when some abnormal behaviors occur, the movements will also be deformed. Based on the rate of change, it can be determined whether each movement is deformed. If deformation is found, individual abnormal behaviors can be identified.

[0204] In cases where both side-view and top-view borders exist and are coupled, it's necessary to process the border change rate of one border segment. Simultaneously, the change rate of the coupled border segment is analyzed to determine if the change rates of the two borders correspond. The equation is:

[0205] ;

[0206] in, , , and These represent the rate of change of the length and width of the border in the side view, and the rate of change of the length and width of the border in the top view, respectively. and These represent the lengths of the side view borders when an individual wild animal with unusual behavior makes a move, specifically the time of the action's termination and the time of its initiation. and The widths of the side view borders represent the time when an individual wild animal with abnormal behavior makes a move, the time when the move ends, and the time when the move begins. and The lengths of the top-view borders represent the time when an individual wild animal with unusual behavior makes a movement, the time when the movement ends and the time when the movement begins; and The widths of the top-view borders represent the time when an individual wild animal with unusual behavior makes a move, the time when the move ends, and the time when the move begins. and These represent the time point at which the action ends and the time point at which the action begins when an individual wild animal with unusual behavior makes an action; This represents the threshold value indicating the difference in the rate of change between the overlapping edges of the side view border and the top view border.

[0207] Specifically, regarding the above equations, if it is found that... Greater than If so, it is necessary to further adjust the length parameter of the overlapping edge of the part's border and recalculate the change rate of the part's border.

[0208] In addition, the rate of change of all multi-directional side views is analyzed to obtain the change rate parameters of the part borders in all directions, so as to obtain the final result.

[0209] For the extra top-view border, if it is not coupled with the side-view border, the rate of change of the top-view border can be obtained directly.

[0210] As described in step S170, the purpose of this step is to determine the behavior pattern of the wild animal individual exhibiting abnormal behavior after determining the bounding box change rate. Then, based on the obtained change rate parameter, the current abnormal behavior pattern can be directly determined, and further, the cause of the abnormal behavior can be determined based on the abnormal behavior pattern. Specifically:

[0211] S171. Obtain the number of times the boundary change rate of all parts of the wild animal individual with abnormal behavior occurs, and obtain the frequency of compulsive behavior.

[0212] The purpose of this step is to determine that when a wild animal exhibits abnormal behavior, it does not necessarily mean that the individual wild animal has abnormal behavior. For example, if an even-toed ungulate rubs its horns and abdomen together, and this behavior is considered abnormal, then if it only occurs once in a long period of time, then the behavior is obviously occasional. However, if the behavior occurs frequently, it means that the behavior is abnormal. Therefore, it is necessary to determine the frequency of the relevant behavior in order to determine whether the abnormal behavior is abnormal.

[0213] Specifically, for the frequency of a specific behavior, it is necessary to determine the rate of change of the boundary of the part and analyze the frequency of the rate of change. If the number of occurrences is not less than 2 (obviously, it can also be set to other numbers), and the frequency is high (by setting a preset frequency and comparing the actual frequency with the preset frequency), the behavior performed by the wild animal is set as a compulsive behavior.

[0214] In some embodiments, abnormal behavior of an individual can be directly obtained, and the change rate of the part border can be obtained based on the abnormal behavior. When the change rate of the part border corresponding to a certain action and the set abnormal behavior is found, it is considered that the animal has engaged in compulsive behavior.

[0215] The frequency of changes in the border of a part can also be determined based on the frequency of the same action. When the frequency is high or exceeds the set threshold, the action corresponding to the border change rate of that part is considered to be a compulsive behavior.

[0216] In some embodiments, if an individual is found to be completely occluded by other individuals, i.e., the lateral border of the animal cannot be obtained, only the rate of change of the top-view border is analyzed to determine the frequency of compulsive behaviors.

[0217] S172. Compare the frequency of the compulsive behavior with the preset frequency of the compulsive behavior. If the frequency of the compulsive behavior is not lower than the preset frequency of the specific behavior, it is determined that the wild animal individual with abnormal behavior has abnormal behavior.

[0218] The purpose of this step is to determine whether there is abnormal behavior in wild animal individuals exhibiting unusual behavior, and to lay the foundation for subsequent work on identifying and analyzing the causes of abnormal behavior.

[0219] Among them, a threshold is set for compulsive behavior to obtain the preset frequency of compulsive behavior occurrence, and the measured results are compared to determine whether wild animal individuals with abnormal behavior patterns have abnormal behaviors.

[0220] Among them, the preset frequency of compulsive behavior can be obtained based on the frequency of compulsive behavior corresponding to the abnormal behavior, that is, it can be determined based on the habits of wild animals.

[0221] S173. Sort the frequency of occurrence of the boundary change rate of the parts of the wild animal individuals with abnormal behavior to obtain high-frequency abnormal behaviors.

[0222] The purpose of this step is to identify the most frequent abnormal behaviors among wild animal individuals when such behaviors may be categorized into multiple behaviors. This allows for the identification of the most prominent abnormal behavioral characteristics based on this information.

[0223] Among them, the frequency of occurrence of the same part border change rate is sorted according to the frequency of occurrence of each type.

[0224] Among them, based on the set occurrence frequency threshold, or the number of occurrence frequencies selected for the occurrence frequency, high-frequency abnormal behaviors are selected, and the abnormal behaviors corresponding to the occurrence frequencies obtained are high-frequency abnormal behaviors.

[0225] S174. Compare the severity of abnormal behavior and the causes of abnormal behavior in the wild animal populations with the high frequency of abnormal behavior and the abnormal behavior of the wild animals to obtain the causes of abnormal behavior of the wild animals.

[0226] The purpose of this step is to analyze the causes of abnormal behavior in wild animals based on the information obtained regarding the high frequency of abnormal behaviors and the severity of abnormal behavior patterns.

[0227] This includes setting up a lookup table for the causes of abnormal behavior, which includes the severity of abnormal behavior in wild animals, and establishing the high-frequency abnormal behaviors included under each severity level. For each high-frequency abnormal behavior, a corresponding cause of the abnormal behavior is set.

[0228] For each type of wild animal population exhibiting abnormal behavior, the severity of the abnormal behavior is assigned a corresponding cause for that behavior abnormality.

[0229] After obtaining the severity of the abnormal behavior of a wild animal population, the cause of the abnormal behavior can be obtained directly from the cause lookup table. Furthermore, when high-frequency abnormal behaviors are obtained, the cause of the specific high-frequency abnormal behavior can be obtained.

[0230] In addition, this application also discloses an intelligent analysis system for abnormal animal behavior, such as Figure 2The diagram shown is a schematic of an intelligent analysis system for abnormal animal behavior disclosed in an embodiment of this application. Specifically:

[0231] An intelligent analysis system for abnormal animal behavior includes a static image acquisition system, a dynamic image acquisition system, an animal feature recognition system, a wild animal population identification system for abnormal gait patterns, a wild animal individual identification system for abnormal gait patterns, a wild animal individual part decomposition system for abnormal gait patterns, an abnormal behavior analysis system, and a database. Its features are:

[0232] The static image acquisition system and the animal feature recognition system are connected to acquire information on wild animal populations in normal and abnormal monitoring areas, and send the acquired monitoring information to the animal feature recognition system.

[0233] The animal feature recognition system is connected to the abnormal wildlife population recognition system to obtain wildlife features within the monitoring area;

[0234] The abnormal behavior wildlife population identification system is also connected to the dynamic image acquisition system, and is used to obtain abnormal behavior wildlife population information based on the wildlife characteristics. After obtaining the abnormal behavior wildlife population information, the system sends shooting instructions and monitoring information to the image acquisition system.

[0235] The abnormal behavior wildlife individual identification system is connected to both the abnormal behavior wildlife population identification system and the dynamic image acquisition system to acquire images of the abnormal behavior wildlife individuals.

[0236] The system for decomposing the parts of wild animals with abnormal behavior patterns and the system for identifying wild animals with abnormal behavior patterns are connected to obtain the bounding boxes of the parts of wild animals with abnormal behavior patterns.

[0237] The abnormal behavior analysis system is connected to the abnormal behavior individual part decomposition system and the abnormal behavior population identification system for wild animals, and is used to obtain the types and causes of abnormal behaviors of the abnormal behavior population and the abnormal behavior of the abnormal behavior individuals of wild animals.

[0238] The abnormal behavior analysis system is also connected to the database, which stores the rate of change of the body part borders of the wild animal individuals with abnormal behavior, the severity of the abnormal behavior of the wild animal population with abnormal behavior, and the correspondence between the causes of the abnormal behavior of wild animals.

[0239] Also includes:

[0240] The static image acquisition system includes a normal monitoring area and an abnormal monitoring area. The normal monitoring area is the existing image acquisition area, and the abnormal monitoring area is other image acquisition areas defined based on the existing image acquisition area.

[0241] In both the normal monitoring area and the abnormal monitoring area, multiple image acquisition devices are set along the edge line of the monitoring area.

[0242] The beneficial effects of this application include:

[0243] 1. Reduced implementation costs for analyzing abnormal wildlife behavior. The technical solution of this application establishes static and dynamic image acquisition systems. The static image acquisition system fully utilizes existing image acquisition systems and selects abnormal monitoring areas based on these systems, establishing corresponding image acquisition systems to effectively utilize existing resources. Simultaneously, the dynamic image acquisition system can autonomously reach the monitoring area to acquire images without fixed setup, allowing for the recycling and reuse of the dynamic image acquisition system, thus reducing the overall system construction and implementation costs.

[0244] 2. It broadens the scope of analysis of abnormal behavior in wild animals. The technical solution of this application uses a static image acquisition system to obtain data on wild animal populations exhibiting abnormal behavior. Then, based on the combined application of static and dynamic image acquisition systems, the abnormal behavior of individual wild animals exhibiting abnormal behavior is analyzed. This achieves joint analysis of abnormal behavior in both populations and individuals, thus broadening the coverage of abnormal behavior analysis.

[0245] 3. Reduced resource consumption in analyzing abnormal behavior of wild animals. The technical solution of this application does not employ image recognition-based appearance analysis for analyzing abnormal behavior of wild animals. Instead, it sets bounding boxes around individual wild animals and analyzes abnormal behavior based on the rate of change of these bounding boxes and the frequency of their occurrence. Compared to image recognition and related algorithms, this significantly reduces resource consumption while ensuring analytical accuracy.

[0246] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to computer program instructions. The aforementioned computer program can be stored in a non-volatile storage medium, and when executed, it performs the steps of the above method embodiments. Alternatively, if the integrated unit of the present invention is implemented as a software functional module and sold or used as an independent product, it can also be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the embodiments of the present invention, or the part that contributes 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 cause an electronic device (which may be a personal computer, server, network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention.

[0247] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. A method for intelligent analysis of abnormal behaviors in wild animals, characterized in that, The method includes: Based on a static image acquisition system, identify the types of wild animal populations entering the monitored area; Based on the aforementioned wildlife population types, obtain wildlife populations exhibiting unusual behavior patterns; Based on the static image acquisition system, a multi-directional side view of the wild animal population with abnormal behavior is obtained; A top-down view of the abnormally behaving wild animal population was obtained using a dynamic image acquisition system. Based on the top view of the wildlife population with abnormal behavior, identify individual wildlife with abnormal behavior. Setting part borders for the wild animal individuals exhibiting abnormal behavior, and obtaining the change rate of the part borders based on the part borders, including: Obtain the body parts of the wild animal individual with abnormal behavior, and set the side view borders on the multi-directional side view; Obtain a top view of the individual wild animal exhibiting abnormal behavior, and set a top view border based on the body parts. The side view border and the top view border of the wild animal individual with abnormal behavior are coupled to obtain the part border; Obtain the rate of change of all edges of the bounding box of the specified region to obtain the rate of change of the bounding box of the region. Based on the change rate of the border of the aforementioned region, the abnormal behavior and causes of wild animals are obtained.

2. The intelligent analysis method for abnormal wildlife behavior according to claim 1, characterized in that, The static image acquisition system identifies the types of wild animal populations entering the monitored area, including: Obtain the types of wild animal populations in the region and the normal distribution areas of wild animal populations to obtain the normal monitoring areas; Based on the normal distribution area, the abnormal distribution area of ​​the wild animal population is obtained; Based on the distribution distance between the normal distribution area and the abnormal distribution area, the distribution area anomaly degree of the wild animal population is obtained; A static image acquisition system is set up in the abnormal distribution area to obtain the abnormal monitoring area; The system acquires the entry status of wild animals in the normal monitoring area and the abnormal monitoring area, and identifies the wild animal population type.

3. The intelligent analysis method for abnormal wildlife behavior according to claim 1, characterized in that, The process of identifying wildlife populations exhibiting unusual behavior patterns based on the aforementioned wildlife population type includes: Set the types of wild animal populations that normally enter the normal monitoring area and obtain the types of normal wild animal populations; The types of wild animal populations entering the normal monitoring area are compared with the types of normal wild animal populations. If they are different, the wild animal populations entering the normal monitoring area are wild animal populations with abnormal behavior patterns, and the wild animal populations with abnormal behavior patterns in the normal monitoring area are obtained. Obtain the types of wild animal populations within the abnormal monitoring area, and obtain wild animal populations exhibiting abnormal behavior patterns within the abnormal monitoring area; Both the abnormal wildlife populations in the normal monitoring area and the abnormal wildlife populations in the abnormal monitoring area are set as abnormal wildlife populations. Identify the types of wildlife populations exhibiting abnormal behavior and determine the severity of such abnormal behavior.

4. The intelligent analysis method for abnormal wildlife behavior according to claim 1, characterized in that, The process of acquiring multi-directional side views of the abnormally behaving wild animal population based on the static image acquisition system includes: Based on the static image acquisition system, multi-directional frame images of the wildlife population with the tracks are acquired; Based on the multi-directional frame image and the orientation of all image acquisition devices in the static image acquisition system, a multi-directional side view is obtained.

5. The intelligent analysis method for abnormal wildlife behavior according to claim 1, characterized in that, The method of acquiring a top-down view of the abnormally behaving wild animal population based on the dynamic image acquisition system includes: After the static image acquisition system acquires information about a wild animal population exhibiting abnormal behavior, the dynamic image acquisition system moves to the corresponding monitoring area. The dynamic image acquisition system moves above the abnormal wildlife population and continuously acquires a top-down view of the abnormal wildlife population.

6. The intelligent analysis method for abnormal wildlife behavior according to claim 1, characterized in that, The method of identifying individual wild animals exhibiting unusual behavior based on a top-view view of the population includes: Based on the top view of the abnormal behavior of the wild animal population, the wild animals in the abnormal behavior population are labeled to obtain individual labeling information; Based on the top view of the wild animal population with abnormal behavior, the individual labeling information is associated with the top view of the wild animal individual with abnormal behavior to obtain information about the wild animal individual with abnormal behavior.

7. The intelligent analysis method for abnormal wildlife behavior according to claim 1, characterized in that, The process of obtaining abnormal behavior and causes of wild animals based on the change rate of the border of the described location includes: The frequency of compulsive behavior is obtained by acquiring the number of times the bounding box change rate of all parts of the wild animal individual with abnormal behavior; The frequency of the compulsive behavior is compared with the preset frequency of the compulsive behavior. If the frequency of the compulsive behavior is not lower than the preset frequency of the specific behavior, it is determined that the wild animal individual with abnormal behavior has abnormal behavior. The frequency of the change rate of the boundary of the part of the wild animal individual with abnormal behavior is sorted to obtain the high-frequency abnormal behavior. By comparing the severity of abnormal behavior and the causes of abnormal behavior in the wild animal populations with the frequency of abnormal behavior and the abnormal behavior of the wild animals, the causes of abnormal behavior in wild animals can be obtained.

8. An intelligent analysis system for abnormal wildlife behavior, used to execute the intelligent analysis method for abnormal wildlife behavior as described in any one of claims 1 to 7, comprising a static image acquisition system, a dynamic image acquisition system, an animal feature recognition system, a wildlife population identification system for abnormal behavior patterns, a wildlife individual identification system for abnormal behavior patterns, a wildlife individual part decomposition system for abnormal behavior patterns, an abnormal behavior analysis system, and a database, characterized in that: The static image acquisition system and the animal feature recognition system are connected to acquire information on wild animal populations in normal and abnormal monitoring areas, and send the acquired monitoring information to the animal feature recognition system. The animal feature recognition system is connected to the abnormal wildlife population recognition system to obtain wildlife features within the monitoring area; The abnormal behavior wildlife population identification system is also connected to the dynamic image acquisition system, and is used to obtain abnormal behavior wildlife population information based on the wildlife characteristics. After obtaining the abnormal behavior wildlife population information, the system sends shooting instructions and monitoring information to the image acquisition system. The abnormal behavior wildlife individual identification system is connected to both the abnormal behavior wildlife population identification system and the dynamic image acquisition system to acquire images of the abnormal behavior wildlife individuals. The system for decomposing the parts of wild animals with abnormal behavior patterns and the system for identifying wild animals with abnormal behavior patterns are connected to obtain the bounding boxes of the parts of wild animals with abnormal behavior patterns. The abnormal behavior analysis system is connected to the abnormal behavior individual part decomposition system and the abnormal behavior population identification system for wild animals, and is used to obtain the types and causes of abnormal behaviors of the abnormal behavior population and the abnormal behavior of the abnormal behavior individuals of wild animals. The abnormal behavior analysis system is also connected to the database, which stores the rate of change of the body part borders of the wild animal individuals with abnormal behavior, the severity of the abnormal behavior of the wild animal population with abnormal behavior, and the correspondence between the causes of the abnormal behavior of wild animals.

9. The intelligent analysis system for abnormal wildlife behavior according to claim 8, characterized in that, The static image acquisition system also includes: The static image acquisition system includes a normal monitoring area and an abnormal monitoring area. The normal monitoring area is the existing image acquisition area, and the abnormal monitoring area is other image acquisition areas defined based on the existing image acquisition area. In both the normal monitoring area and the abnormal monitoring area, multiple image acquisition devices are set along the edge line of the monitoring area.

Images