CN113963298A - Wild animal identification tracking and behavior detection system, method, equipment and storage medium based on computer vision - Google Patents
Wild animal identification tracking and behavior detection system, method, equipment and storage medium based on computer vision Download PDFInfo
- Publication number
- CN113963298A CN113963298A CN202111239920.5A CN202111239920A CN113963298A CN 113963298 A CN113963298 A CN 113963298A CN 202111239920 A CN202111239920 A CN 202111239920A CN 113963298 A CN113963298 A CN 113963298A
- Authority
- CN
- China
- Prior art keywords
- animal
- behavior
- tracking
- target
- wild
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 241001465754 Metazoa Species 0.000 title claims abstract description 178
- 238000001514 detection method Methods 0.000 title claims abstract description 70
- 238000000034 method Methods 0.000 title claims abstract description 53
- 230000004438 eyesight Effects 0.000 title claims abstract description 29
- 238000003860 storage Methods 0.000 title claims abstract description 17
- 238000004458 analytical method Methods 0.000 claims abstract description 37
- 238000012545 processing Methods 0.000 claims abstract description 29
- 241000894007 species Species 0.000 claims abstract description 26
- 230000006399 behavior Effects 0.000 claims description 94
- 230000033001 locomotion Effects 0.000 claims description 29
- 238000012549 training Methods 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 15
- 230000000694 effects Effects 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 14
- 230000009471 action Effects 0.000 claims description 12
- 239000000203 mixture Substances 0.000 claims description 12
- 230000000877 morphologic effect Effects 0.000 claims description 12
- 230000002159 abnormal effect Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 9
- 230000010354 integration Effects 0.000 claims description 9
- 230000003542 behavioural effect Effects 0.000 claims description 8
- 238000013135 deep learning Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 7
- 230000008520 organization Effects 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000003628 erosive effect Effects 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 9
- 238000009826 distribution Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 241000282994 Cervidae Species 0.000 description 3
- 241000282376 Panthera tigris Species 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000009395 breeding Methods 0.000 description 3
- 230000001488 breeding effect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000013011 mating Effects 0.000 description 3
- 241000282324 Felis Species 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000004297 night vision Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000032696 parturition Effects 0.000 description 2
- 230000000284 resting effect Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 206010000117 Abnormal behaviour Diseases 0.000 description 1
- 241000282979 Alces alces Species 0.000 description 1
- 101100476202 Caenorhabditis elegans mog-2 gene Proteins 0.000 description 1
- 241000282461 Canis lupus Species 0.000 description 1
- 241001481760 Erethizon dorsatum Species 0.000 description 1
- 241000283899 Gazella Species 0.000 description 1
- 241000282412 Homo Species 0.000 description 1
- 241000406668 Loxodonta cyclotis Species 0.000 description 1
- 206010042008 Stereotypy Diseases 0.000 description 1
- 241000282458 Ursus sp. Species 0.000 description 1
- 241001464837 Viridiplantae Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000010171 animal model Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000669 biting effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000009193 crawling Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000012173 estrus Effects 0.000 description 1
- 230000035558 fertility Effects 0.000 description 1
- 230000037406 food intake Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007758 mating behavior Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000024121 nodulation Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008560 physiological behavior Effects 0.000 description 1
- 230000004962 physiological condition Effects 0.000 description 1
- 230000035790 physiological processes and functions Effects 0.000 description 1
- 244000062645 predators Species 0.000 description 1
- 230000010539 reproductive behavior Effects 0.000 description 1
- 230000001850 reproductive effect Effects 0.000 description 1
- 230000009183 running Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000011410 subtraction method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
A wild animal identification tracking and behavior detection system, a method, equipment and a storage medium based on computer vision belong to the field of identification tracking and behavior detection systems and methods, and are provided for solving the problems that wild animals cannot be effectively monitored in harsh environments, manual supervision cannot track the wild animals for 24 hours, and the manual processing capacity of a large amount of data is limited; the problem that wild animals cannot be effectively monitored manually in harsh field conditions is solved, the state analysis of the trace recognition species of the wild animals can be found in time, signals are sent in time, and the problem that 24-hour tracking recording is difficult to realize through manual supervision is solved.
Description
Technical Field
The invention relates to the field of identification, tracking and behavior detection systems and methods, in particular to a wild animal identification, tracking and behavior detection system, method, equipment and storage medium based on computer vision.
Background
Under the global wild animal protection environment, a large number of cameras are densely distributed in natural ecological regions of natural ecological research institutions and wild animal protection organizations at home and abroad, wild animals in a monitoring range are shot and captured, and image data are acquired.
In the prior art, behavior identification is carried out on running, resting, ingestion, biting and the like of the captive porcupine, and a high monitoring accuracy is obtained through a mixed Gaussian background model, ORB feature point detection and a data mining classification algorithm. However, at present, the related research work on domestic wild animal populations is less.
The method has the problems that wild animals cannot be effectively monitored under severe conditions, manual supervision cannot be carried out for 24 hours, and the manual processing capacity of a large amount of data is limited.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides a wild animal identification tracking and behavior detection system, method, equipment and storage medium based on computer vision, which can solve the problems that the wild animal cannot be effectively monitored in a severe environment, manual supervision cannot be carried out for 24 hours, and the manual processing capacity of a large amount of data is limited. Compared with the existing mode, the method has the advantages of low labor cost, high speed and strong environmental adaptability.
The technical scheme adopted by the invention is as follows:
the wild animal identification tracking and behavior detection method based on computer vision comprises the following steps:
s1, carrying out video acquisition on wild animals;
step S2, preprocessing the collected video stream data;
step S3, aiming at the data processed in the step S2, carrying out target identification and recording on the wild animals in the video picture information;
s4, classifying the result information of the wild animal target identification and record in the S3, and formatting and integrating;
s5, training various animal movement joint node models by using the wild animal target identification and recording result information classified and integrated in the S4, compiling specific data parameters according to animal behavioral knowledge of the species, and building a specific behavioral parameter information database for each wild animal;
step S6, physical parameters of the posture and the action behavior of the animal are obtained according to the detection and analysis of the limb model;
step S7, comparing the physical parameters of the animal body state and the action behavior obtained in the step S6 with the animal specific characteristic behavior database obtained in the step S5 to obtain the animal state and behavior explanation;
and step S8, carrying out classification statistics on the analysis result and uploading for recording.
Further, the method also comprises a step S9 of utilizing the analysis result obtained in the step S8 to judge whether the animal behavior is in an abnormal state, if so, giving a warning and feeding back to the animal protection organization personnel, and if not, feeding back information to the step S5 for updating the animal behavior parameter information database.
Further, in step S2, the specific method for preprocessing the collected video stream data is as follows:
step S21, frame extraction is carried out on the video stream to be converted into picture information, pixel points are taken to carry out gray scale conversion on the picture by using a weighted average method, Gaussian filtering is used to generate a weight matrix, filtering processing is carried out on image interference signals and noise points, and after repeated operation is carried out on all the pixel points, an image after Gaussian blur is obtained;
s22, aiming at the image after the Gaussian blur obtained in the step S21, the image is enhanced, the edge sharpening is carried out by combining a gradient method, then the line binarization processing is carried out on the image by adopting a triangular threshold segmentation algorithm, the gray value of a pixel point on the image is set to be 0 or 255, so that the whole image presents an obvious black and white effect, and the outline of a target is highlighted;
and step S23, performing morphological filtering processing on the picture, subtracting the expansion map and the erosion map by using a morphological gradient algorithm, highlighting the edge of the image after operating the binary image, and keeping the edge contour of the target by using a morphological gradient.
Further, in step S3, the specific method for identifying the target of the wild animal in the video picture information is as follows:
s31, modeling through a Gaussian mixture model, separating the background before object motion, detecting the moving object, processing the image of the moving object, and drawing a minimum circumscribed rectangle to realize the segmentation and detection of the moving object;
and step S32, extracting the target by a frame difference method by using a background extraction algorithm with the frame difference method and the Gaussian mixture model fused, updating the background by using the Gaussian mixture background model, subtracting the intermediate frame image and the background image to extract the target, carrying out AND operation on the two detected targets by motion strategy analysis, and obtaining the final motion target by connectivity detection and morphological processing.
Further, in step S5, a specific method for constructing a specific behavior parameter information database for each wild animal is to use a Yolov4 algorithm, and perform deep learning training after using a large number of video and picture data sets of various wild animals and labeling the types and positions of the various animals therein, so as to obtain a training model with higher precision.
Further, the physical parameters of the body state and the action behavior of the animal obtained by the detection and analysis of the limb model in step S6 are specifically: using ANY ANY-maze software, according to the shape (circle or rectangle) set by experimenters, the shot animal moving images are transmitted to an analysis computer, and the position, speed, residence time, moving track and moving distance parameters of one or more animals in one or more areas at different time are recorded.
Wild animal discernment tracking and action detecting system based on computer vision, including target detection and tracking platform, behavioristics analysis platform, wherein:
the target detection and tracking platform comprises a video stream preprocessing module, a motion detection and tracking module and a target tracking identification and recording module, wherein the video stream preprocessing module is used for preprocessing acquired video stream data;
the motion detection and tracking module is used for carrying out motion detection and tracking on the processed video stream data and capturing and detecting a target;
the target tracking, identifying and recording module is used for tracking each animal, identifying species of various occurring animals and recording the species and the track of each animal;
the behavioristics analysis platform comprises an information classification and format integration module, a wild animal behavior training module and a wild animal state and behavior comparison module;
the information classification and format integration module is used for receiving species and track result information of each animal recorded by a target tracking identification and recording module in the target detection and tracking platform, classifying the wild animal target identification and recorded result information, and performing formatting integration;
the wild animal behavior training module compiles a specific data parameter according to the animal behavior knowledge of the species, and builds a specific behavior parameter information database for each wild animal;
and the wild animal state and behavior comparison module is used for comparing the obtained behavior parameters of the wild animals with the established behavior parameter information database to obtain the animal state and behavior explanation.
The wild animal identification tracking and behavior detection system based on computer vision comprises a behavioristics analysis platform and an early warning module, wherein the early warning module is used for judging whether the animal behavior is in an abnormal state or not, and giving a warning and feeding the warning back to an animal protection organization personnel if the animal behavior is abnormal.
Computer vision based wildlife identification tracking and behavior detection apparatus, comprising a memory storing a computer program and a processor, wherein the processor when executing the computer program implements the steps of the computer vision based wildlife identification tracking and behavior detection method as described in any one of the preceding claims.
A computer-readable storage medium, on which a computer program is stored, wherein the computer program, when executed by a processor, implements a computer vision based wildlife identification tracking and behavior detection method as described in any one of the preceding.
The invention has the beneficial effects that:
1. the method and the device realize high-precision identification and classification of animals appearing in real-time video streams or independent videos of the camera. The moving track of each animal is tracked and analyzed corresponding to the independent database of each animal, the problem that wild animals cannot be effectively monitored manually in an environment with harsh field conditions is solved, the trace of the wild animals can be found in time, the species state can be identified, and signals can be sent in time.
2. Model training and deep learning are carried out on each wild animal by using a large amount of data sets of various wild animals, all-weather physiological state and behavior track detection of species of animals in a field environment or a zoo is realized for the acquired video stream according to visual neural network analysis and dynamic image processing, and data of intelligent analysis is transmitted and recorded in real time, so that the problems that 24-hour tracking and recording are difficult to realize by manual supervision are solved, and a large amount of manpower and material resources are saved.
3. The deep learning training model uses a large number of data sets, the high-efficiency video stream processing program is combined with the high-speed computing capability and specific algorithm optimization of a computer system, the problems of limited manual processing capability and short observation time are solved, the mechanized manual work is reduced, the working efficiency is greatly improved, and the human resources are saved.
4. The camera and the video analysis system support the infrared night vision processing function, and the problem that manual night vision difference detection and observation are difficult is solved.
Drawings
FIG. 1 is a block diagram of a flow chart of an implementation of a computer vision-based wildlife identification tracking and behavior detection system;
FIG. 2 is a technical flow diagram of a computer vision based wildlife identification tracking and behavior detection system;
FIG. 3 is a block diagram of a video detection system for a computer vision based wildlife identification tracking and behavior detection system;
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
The wild animal identification tracking and behavior detection system, method, equipment and storage medium based on computer vision can be applied to the following application scenes:
1. in the aspect of scientific research:
(1) application scenarios: areas in the nature reserve where wild animals often emerge (such as at plaque edges or galleries);
(2) the application method comprises the following steps: the monitoring system is arranged in an area where wild animals usually appear in a natural protection area, and the full-automatic 24-hour behavior detection is carried out on the animals appearing in the monitoring range of the protection area and the curtain of the animal activity area.
(3) The application effect is as follows:
firstly, first-hand data of animal behaviors can be obtained under the condition of saving manpower and time to the maximum extent;
and secondly, providing sufficient data for basic research in animal science fields such as behavior ecology, physiological conditions, life habits and the like of the target animal in the monitoring range.
2. And (4) safety monitoring:
(1) application scenarios: residential areas where wildlife and humans may conflict (e.g., rural areas near natural conservation);
(2) the application method comprises the following steps: the safety monitoring of the residential area is combined with the system, and the animal presence and absence condition and the movement track of the monitored area are monitored.
(3) The application effect is as follows:
the early warning method can give timely early warning to the types, the quantity and the behaviors of wild animals which are about to appear in human activity areas;
the first time avoids the positive conflict between human and wild animals as far as possible.
3. Animal domestication and breeding aspects:
(1) application scenarios: zoos, animal farms, and animal care farms (e.g., horse farms, deer farms, etc.);
(2) the application method comprises the following steps: the system is applied to the raising environment to track and analyze the activity condition and special period behaviors (such as spouse, parturition and the like) of the target animal.
(3) The application effect is as follows:
firstly, the technology is applied to zoos or animal farms, so that the reproductive rate and the fertility of animals can be reasonably and efficiently predicted;
secondly, labor force for a large amount of manual observation in breeding seasons of a breeding farm is replaced, and the influence of the human activities on animals such as frightening and stress is reduced to the minimum while the state of the bred animals is monitored.
4. The field of laboratory animal research:
(1) application scenarios: laboratories and the like;
(2) the application method comprises the following steps: the monitoring module is well built in the culture environment of the animal, so that 24-hour detection of the specific animal can be realized, the behavior state of the animal, the frequency, the times and the occupied time of the specific behavior all day or any time period of the behavior are automatically identified and recorded, and various data analysis is carried out.
(3) The application effect is as follows: manual supervision and recording are not needed, and manpower and material resources are saved.
5. Large scale target detection field:
(1) application scenarios: large-scale target detection, etc.;
(2) the application method comprises the following steps: recording the situations of insufficient manual monitoring energy and difficult manual monitoring of animals in dark environments and other harsh environmental requirements.
(3) The application effect is as follows: manpower and material resources are saved, and the difficulty of work is reduced.
As shown in FIG. 1 and FIG. 2, the implementation method of the wild animal identification, tracking and behavior detection system based on computer vision is as follows:
(1) preprocessing a video:
firstly, frame extraction is carried out on video stream and converted into picture information, pixel points are taken to carry out gray scale conversion on the picture by using a weighted average method, Gaussian filtering is used for generating a weight matrix to carry out filtering processing on image interference signals and noise points, and the image after Gaussian blur can be obtained after repeated operation is carried out on all the points.
And after processing, enhancing the picture and carrying out edge sharpening by combining a gradient method. Then, taking Triangle to the image
The (triangular threshold segmentation) algorithm carries out binarization processing, namely the gray value of a pixel point on the image is set to be 0 or 255, so that the whole image presents obvious black and white effect, and the data volume in the image is greatly reduced after binarization, thereby highlighting the outline of a target.
Finally, morphological filtering processing is carried out on the picture, a morphological gradient algorithm is used for subtracting the expansion map and the erosion map, the edge can be highlighted after the binary image is operated, and the edge outline of the object is reserved by using the morphological gradient.
(2) Motion detection and tracking:
the method comprises the steps of carrying out object motion front background separation by using MOG2 (Gaussian mixture model) to detect a moving object, finally carrying out image processing on the object, drawing a minimum circumscribed rectangle of the object, and achieving segmentation and detection on the moving object.
The grey histogram of an image reflects the frequency of occurrence of a certain grey value in the image, i.e. an estimate of the probability density of the grey of the image. If the target area contained in the image is larger than the background area, and the background area and the target area have certain difference in gray scale, the gray scale histogram of the image has a double peak-valley shape, wherein one peak corresponds to the target, and the other peak corresponds to the central gray scale of the background. For complex multimodal images, the segmentation problem of the image can be solved by considering the multimodal characteristic of the histogram as superposition of multiple gaussian distributions.
Foreground means that any meaningful moving object is the foreground under the assumption that the background is stationary.
The gaussian model is a model formed based on a gaussian probability density function (normal distribution curve) by accurately quantizing an object using the gaussian probability density function (normal distribution curve) and decomposing one object into a plurality of objects. The R, G, B three-channel pixel value variation for each pixel is characterized by a mixed gaussian model distribution. This has the advantage that multiple modes of pixel value variation (e.g. water ripples, sloshing leaves, etc.) can be present at the same pixel location.
The Gaussian mixture model uses n Gaussian models to represent the characteristics of each pixel point in the image, the Gaussian mixture model is updated after a new frame of image is obtained, each pixel point in the current image is matched with the Gaussian mixture model, if the matching is successful, the point is judged to be a background point, and if the matching is not successful, the point is judged to be a foreground point. The whole Gaussian model is mainly determined by two parameters, namely variance and mean, and the stability, accuracy and convergence of the model are directly influenced by learning the mean and the variance and adopting different learning mechanisms. Since we model the background extraction of moving objects, it is necessary to update both the variance and mean parameters in the gaussian model in real time. In order to improve the learning capability of the model, the improved method adopts different learning rates for updating the mean value and the variance; in order to improve the detection effect of a large and slow moving target in a busy scene, the concept of weight mean value is introduced, a background image is established and updated in real time, and then the classification of foreground and background is carried out on pixel points by combining the weight, the weight mean value and the background image.
It selects an appropriate number of gaussian distributions for each pixel, which can better accommodate lighting variations of different scenes, etc. The algorithm content is as follows: and a background extraction algorithm fusing a frame difference method and a Gaussian mixture model. Extracting a target by a frame difference method, updating a background by using Gaussian mixture background modeling, performing difference between an intermediate frame image and a background image, extracting the target, performing AND operation on two detected targets by motion strategy analysis, and obtaining a final motion target by connectivity detection and morphological processing.
And carrying out differential operation on two or three continuous frames of images in time, subtracting pixels corresponding to different frames, judging the absolute value of the gray difference, and judging the moving target when the absolute value exceeds a certain threshold value, thereby realizing the detection function of the target.
The background subtraction method is to approximate the pixel value of the background image by using a parameter model of the background, and to perform differential comparison between the current frame and the background image to realize the detection of the motion region, wherein the pixel region with larger difference is regarded as the motion region, and the pixel region with smaller difference is regarded as the background region, so as to realize the background modeling and the updating thereof.
(3) A target identification tracking and behavior analysis system platform:
the digital image processing is carried out on the preprocessed shooting data, the system automatically identifies the species according to the content captured by the video, a motion model can be established for the motion track of the animal according to every several frames and the next frame, and the establishing process comprises the following steps: and (3) using a Yolov4 algorithm, using a GPU through Cuda and Opencvs, using a large amount of video and picture data sets of various wild animals, labeling the types and positions of various animals, and then performing deep learning training and detection on the model to obtain a high-precision training model. And the Yolov4 is called to process the video collected by the scene on the spot, so that the automatic capture and identification of various animals can be realized.
The deep learning training process comprises the following steps: based on a Yolov4 algorithm, a large number of data sets of various animals are used for carrying out position positioning on different animals to label classified types, deep learning is carried out to train training models of various animals, and animal type identification and dynamic capture of real-time collected video data are realized.
The position of the next frame of animal is predicted through the model, so that the tracking efficiency (target tracking) is improved, an independent data set is established for each animal (such as northeast tiger, wolf, bear and the like) through a large number of sampling, and respective model libraries are trained.
In this case, the moving area of the animal can be recognized by ANY software such as ANY ANY-maze, such as a circle, a rectangle, or the like, set by the experimenter. The moving images of the animals are transmitted to an analysis computer, and the parameters of interest of researchers, such as the positions, speeds, residence times, movement tracks, movement distances and the like of one or more animals in one or more areas at different times can be recorded. The software automatically classifies and counts the parameters according to the design of researchers to obtain the report of the animal activity condition. The method can refer to a DeepLabCut algorithm to carry out limb nodulation capture and tracking on animals. And performing nodal tracking sampling on the joints of the animal by means of DeepLabCut to obtain more detailed data about the posture of the animal. And a behavior analysis method based on motion (locotion) and a shape analysis (shape analysis) method, wherein the barycentric position (expressed in X, Y coordinates) of the object is obtained, the barycentric coordinate position is used as the position of the animal, and then parameters such as speed, distance, trajectory, etc. can be calculated by using a series of central positions of several frames before and after the barycentric coordinate position. For shape recognition methods such as: extracting boundary characteristics, wavelet transformation, regional decomposition, Fourier descriptor and the like.
Wild animals in the video image data are captured, species identification is carried out on the wild animals, and action tracks of the wild animals are marked and recorded. And compiling specific data parameters of the identified species result according to animal behavioral knowledge of the species, and then mapping the specific data parameters to a behavioral parameter information base which is compiled in advance. The specific data parameters are as follows: based on animal ethology, PAE behavioral profile: animal behavior was studied by analysis of various parts of the disassembled animal body. The method comprises the following steps: resting, walking, feeding, approaching, stereotypy, etc. Most notably, reproductive behavior (e.g., puppetry/mating/parturition, etc.) of various animals such as elk: the male deer horn is pushed against a pile of fresh green plants to dazzle in front of the female deer group, the fighting behaviors of the two stags are the doll seeking behaviors, the mating behavior with larger actions is groveling, and the amplitude is generally obvious. Felines (tigers) also have obvious crawling behavior before mating, and are generally female tigers. The mating period, the feeding time, the exercise time and the rest time of the animals and the proportion of the animals in the non-estrus period are greatly different. Anti-predation behavior of gazelle: jump in place to show the predator his or her ability to escape. Engraving behavior: plucking hair, pacing large feline (regularly stepping on the same route repeatedly), winding (bear and elephant).
Wild animal discernment tracking and action detecting system based on computer vision, including target detection and tracking platform, behavioristics analysis platform, wherein:
the target detection and tracking platform comprises a video stream preprocessing module, a motion detection and tracking module and a target tracking identification and recording module, wherein the video stream preprocessing module is used for preprocessing acquired video stream data;
the motion detection and tracking module is used for carrying out motion detection and tracking on the processed video stream data and capturing and detecting a target;
the target tracking, identifying and recording module is used for tracking each animal, identifying species of various occurring animals and recording the species and the track of each animal;
the behavioristics analysis platform comprises an information classification and format integration module, a wild animal behavior training module and a wild animal state and behavior comparison module;
the information classification and format integration module is used for receiving species and track result information of each animal recorded by a target tracking identification and recording module in the target detection and tracking platform, classifying the wild animal target identification and recorded result information, and performing formatting integration;
the wild animal behavior training module compiles a specific data parameter according to the animal behavior knowledge of the species, and builds a specific behavior parameter information database for each wild animal;
and the wild animal state and behavior comparison module is used for comparing the obtained behavior parameters of the wild animals with the established behavior parameter information database to obtain the animal state and behavior explanation.
As shown in FIG. 3, the wild animal identification, tracking and behavior detection system based on computer vision comprises a video detection system which is used as an early warning module in the system, wherein the early warning module is used for judging whether the animal behavior is in an abnormal state, and if the animal behavior is abnormal, giving a warning and feeding back the warning to the animal protection organization personnel.
The video detection system comprises a multi-angle high-definition camera, a multi-angle infrared camera, a switch, a server, a display control center and storage equipment. The method comprises the technical links of testing and adaptive debugging of a software system under the current environment, transmission and communication of data streams, an ambient light source, power supply and the like.
And (4) carrying out one-to-one accurate analysis and explanation and recording on the physical behavior of the species by utilizing the established behavioristics analysis platform. Meanwhile, the result analyzed by the behavioristics analysis platform is fed back to the recognition and tracking platform, and the recognition and tracking platform can reflect the feedback result to the more detailed limb nodularization tracking of the animal under the state, so that the captured data is more precise, a positive feedback high-precision analysis processing system is realized, and the result reliability and objectivity are greatly improved.
And mapping the model to a model corresponding to the animal of the species, performing specific analysis processing on the physical state of the species, and performing behavior state analysis on the species by combining with each pre-written species characteristic library (physical state, preference, habit and the like), wherein if abnormal behaviors are found, the function of alarming and prompting a worker to process in time is realized.
The present embodiments may be provided as a method, system, or computer program product by those skilled in the art using the systems and methods mentioned in the foregoing embodiments. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects, or a combination of both. Furthermore, the present embodiments may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
A flowchart or block diagram of a method, apparatus (system), and computer program product according to the present embodiments is depicted. It will be understood that each flow or block of the flowchart illustrations or block diagrams, and combinations of flows or blocks in the flowchart illustrations or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows, or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
Claims (10)
1. The wild animal identification tracking and behavior detection method based on computer vision is characterized by comprising the following steps of:
s1, carrying out video acquisition on wild animals;
step S2, preprocessing the collected video stream data;
step S3, aiming at the data processed in the step S2, carrying out target identification and recording on the wild animals in the video picture information;
s4, classifying the result information of the wild animal target identification and record in the S3, and formatting and integrating;
s5, training various animal movement joint node models by using the wild animal target identification and recording result information classified and integrated in the S4, compiling specific data parameters according to animal behavioral knowledge of the species, and building a specific behavioral parameter information database for each wild animal;
step S6, physical parameters of the posture and the action behavior of the animal are obtained according to the detection and analysis of the limb model;
step S7, comparing the physical parameters of the animal body state and the action behavior obtained in the step S6 with the animal specific characteristic behavior database obtained in the step S5 to obtain the animal state and behavior explanation;
and step S8, carrying out classification statistics on the analysis result and uploading for recording.
2. The computer vision based wildanimal identification tracking and behavior detection method as claimed in claim 1, wherein: and step S9, judging whether the animal behavior is in an abnormal state by using the analysis result obtained in the step S8, giving a warning and feeding back to the animal protection organization personnel if the animal behavior is in an abnormal state, and feeding back information to the step S5 for updating the animal behavior parameter information database if the animal behavior is not in an abnormal state.
3. The computer vision based wildanimal identification tracking and behavior detection method as claimed in claim 1, wherein: in step S2, the specific method for preprocessing the collected video stream data is as follows:
step S21, frame extraction is carried out on the video stream to be converted into picture information, pixel points are taken to carry out gray scale conversion on the picture by using a weighted average method, Gaussian filtering is used to generate a weight matrix, filtering processing is carried out on image interference signals and noise points, and after repeated operation is carried out on all the pixel points, an image after Gaussian blur is obtained;
s22, aiming at the image after the Gaussian blur obtained in the step S21, the image is enhanced, the edge sharpening is carried out by combining a gradient method, then the line binarization processing is carried out on the image by adopting a triangular threshold segmentation algorithm, the gray value of a pixel point on the image is set to be 0 or 255, so that the whole image presents an obvious black and white effect, and the outline of a target is highlighted;
and step S23, performing morphological filtering processing on the picture, subtracting the expansion map and the erosion map by using a morphological gradient algorithm, highlighting the edge of the image after operating the binary image, and keeping the edge contour of the target by using a morphological gradient.
4. The computer vision based wildanimal identification tracking and behavior detection method as claimed in claim 1, wherein: in step S3, the specific method for identifying the target of the wild animal in the video picture information is as follows:
s31, modeling through a Gaussian mixture model, separating the background before object motion, detecting the moving object, processing the image of the moving object, and drawing a minimum circumscribed rectangle to realize the segmentation and detection of the moving object;
and step S32, extracting the target by a frame difference method by using a background extraction algorithm with the frame difference method and the Gaussian mixture model fused, updating the background by using the Gaussian mixture background model, subtracting the intermediate frame image and the background image to extract the target, carrying out AND operation on the two detected targets by motion strategy analysis, and obtaining the final motion target by connectivity detection and morphological processing.
5. The computer vision based wildanimal identification tracking and behavior detection method as claimed in claim 1, wherein: in the step S5, a specific method for constructing a specific behavior parameter information database for each wild animal is to use a Yolov4 algorithm, and perform deep learning training after using a large number of video and picture data sets of various wild animals and labeling the types and positions of various animals therein, so as to obtain a training model with higher precision.
6. The computer vision based wildanimal identification tracking and behavior detection method as claimed in claim 1, wherein: in step S6, the physical parameters of the posture and the action behavior of the animal obtained by the detection and analysis of the limb model are specifically: using ANY ANY-maze software, according to the shape (circle or rectangle) set by experimenters, the shot animal moving images are transmitted to an analysis computer, and the position, speed, residence time, moving track and moving distance parameters of one or more animals in one or more areas at different time are recorded.
7. Wild animal discernment tracking and action detecting system based on computer vision, its characterized in that: target detection and tracking platform, behavioural analysis platform, wherein:
the target detection and tracking platform comprises a video stream preprocessing module, a motion detection and tracking module and a target tracking identification and recording module, wherein the video stream preprocessing module is used for preprocessing acquired video stream data;
the motion detection and tracking module is used for carrying out motion detection and tracking on the processed video stream data and capturing and detecting a target;
the target tracking, identifying and recording module is used for tracking each animal, identifying species of various occurring animals and recording the species and the track of each animal;
the behavioristics analysis platform comprises an information classification and format integration module, a wild animal behavior training module and a wild animal state and behavior comparison module;
the information classification and format integration module is used for receiving species and track result information of each animal recorded by a target tracking identification and recording module in the target detection and tracking platform, classifying the wild animal target identification and recorded result information, and performing formatting integration;
the wild animal behavior training module compiles a specific data parameter according to the animal behavior knowledge of the species, and builds a specific behavior parameter information database for each wild animal;
and the wild animal state and behavior comparison module is used for comparing the obtained behavior parameters of the wild animals with the established behavior parameter information database to obtain the animal state and behavior explanation.
8. Wild animal discernment tracking and action detecting system based on computer vision, its characterized in that: the behavioristics analysis platform further comprises an early warning module, wherein the early warning module judges whether the animal behaviour is in an abnormal state, and gives a warning and feeds the warning back to the animal protection organization personnel if the animal behaviour is in the abnormal state.
9. Computer vision based wildlife identification, tracking and behavior detection device, comprising a memory storing a computer program and a processor implementing the steps of the computer vision based wildlife identification, tracking and behavior detection method of any one of claims 1 to 6 when the computer program is executed by the processor.
10. A computer-readable storage medium having stored thereon a computer program, characterized in that: the computer program when executed by a processor implements the computer vision based wildlife identification tracking and behavior detection method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111239920.5A CN113963298A (en) | 2021-10-25 | 2021-10-25 | Wild animal identification tracking and behavior detection system, method, equipment and storage medium based on computer vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111239920.5A CN113963298A (en) | 2021-10-25 | 2021-10-25 | Wild animal identification tracking and behavior detection system, method, equipment and storage medium based on computer vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113963298A true CN113963298A (en) | 2022-01-21 |
Family
ID=79466638
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111239920.5A Pending CN113963298A (en) | 2021-10-25 | 2021-10-25 | Wild animal identification tracking and behavior detection system, method, equipment and storage medium based on computer vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113963298A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115861644A (en) * | 2022-11-21 | 2023-03-28 | 广东鉴面智能科技有限公司 | Biodiversity identification and analysis method, system and medium |
| WO2023165299A1 (en) * | 2022-03-03 | 2023-09-07 | 石河子大学 | Intelligent analysis system used for ethology of various high-density tiny insects |
| CN117079397A (en) * | 2023-09-27 | 2023-11-17 | 青海民族大学 | Wild human and animal safety early warning method based on video monitoring |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102509305A (en) * | 2011-09-26 | 2012-06-20 | 浙江工业大学 | Animal behavior detection device based on omnidirectional vision |
| CN104915655A (en) * | 2015-06-15 | 2015-09-16 | 西安电子科技大学 | Multi-path monitor video management method and device |
| CN107507221A (en) * | 2017-07-28 | 2017-12-22 | 天津大学 | With reference to frame difference method and the moving object detection and tracking method of mixed Gauss model |
| CN107886086A (en) * | 2017-12-01 | 2018-04-06 | 中国农业大学 | A kind of target animal detection method and device based on image/video |
| CN111297367A (en) * | 2019-11-26 | 2020-06-19 | 北京海益同展信息科技有限公司 | Animal state monitoring method and device, electronic equipment and storage medium |
| CN111666881A (en) * | 2020-06-08 | 2020-09-15 | 成都大熊猫繁育研究基地 | Giant panda pacing, bamboo eating and oestrus behavior tracking analysis method |
| CN112288771A (en) * | 2020-10-14 | 2021-01-29 | 华南农业大学 | Method for extracting motion tracks of multiple pig bodies and analyzing behaviors in group environment |
| CN112767435A (en) * | 2021-03-17 | 2021-05-07 | 深圳市归位科技有限公司 | Method and device for detecting and tracking captive target animal |
-
2021
- 2021-10-25 CN CN202111239920.5A patent/CN113963298A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102509305A (en) * | 2011-09-26 | 2012-06-20 | 浙江工业大学 | Animal behavior detection device based on omnidirectional vision |
| CN104915655A (en) * | 2015-06-15 | 2015-09-16 | 西安电子科技大学 | Multi-path monitor video management method and device |
| CN107507221A (en) * | 2017-07-28 | 2017-12-22 | 天津大学 | With reference to frame difference method and the moving object detection and tracking method of mixed Gauss model |
| CN107886086A (en) * | 2017-12-01 | 2018-04-06 | 中国农业大学 | A kind of target animal detection method and device based on image/video |
| CN111297367A (en) * | 2019-11-26 | 2020-06-19 | 北京海益同展信息科技有限公司 | Animal state monitoring method and device, electronic equipment and storage medium |
| CN111666881A (en) * | 2020-06-08 | 2020-09-15 | 成都大熊猫繁育研究基地 | Giant panda pacing, bamboo eating and oestrus behavior tracking analysis method |
| CN112288771A (en) * | 2020-10-14 | 2021-01-29 | 华南农业大学 | Method for extracting motion tracks of multiple pig bodies and analyzing behaviors in group environment |
| CN112767435A (en) * | 2021-03-17 | 2021-05-07 | 深圳市归位科技有限公司 | Method and device for detecting and tracking captive target animal |
Non-Patent Citations (2)
| Title |
|---|
| 刘波 等: ""基于四帧帧差和混合高斯模型的运动目标检测"", 《平顶山学院学报》, vol. 32, no. 5, 31 October 2017 (2017-10-31), pages 13 - 16 * |
| 马贺: ""蛋鸡对光照自主选择的偏好性试验研究"", 《中国博士学位论文全文数据库农业科技辑》, no. 03, 15 March 2016 (2016-03-15), pages 6 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023165299A1 (en) * | 2022-03-03 | 2023-09-07 | 石河子大学 | Intelligent analysis system used for ethology of various high-density tiny insects |
| US11967182B2 (en) | 2022-03-03 | 2024-04-23 | Shihezi University | Intelligent analysis system applied to ethology of various kinds of high-density minimal polypides |
| CN115861644A (en) * | 2022-11-21 | 2023-03-28 | 广东鉴面智能科技有限公司 | Biodiversity identification and analysis method, system and medium |
| CN117079397A (en) * | 2023-09-27 | 2023-11-17 | 青海民族大学 | Wild human and animal safety early warning method based on video monitoring |
| CN117079397B (en) * | 2023-09-27 | 2024-03-26 | 青海民族大学 | Wild human and animal safety early warning method based on video monitoring |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113963298A (en) | Wild animal identification tracking and behavior detection system, method, equipment and storage medium based on computer vision | |
| EP3125151A2 (en) | Inventory, growth, and risk prediction using image processing | |
| Balakrishna et al. | Application of IOT and machine learning in crop protection against animal intrusion | |
| CN108229407A (en) | A kind of behavioral value method and system in video analysis | |
| CN102915446A (en) | Plant disease and pest detection method based on SVM (support vector machine) learning | |
| CN112528823B (en) | Method and system for analyzing batcharybus movement behavior based on key frame detection and semantic component segmentation | |
| Noe et al. | Automatic detection and tracking of mounting behavior in cattle using a deep learning-based instance segmentation model | |
| Ratnayake et al. | Towards computer vision and deep learning facilitated pollination monitoring for agriculture | |
| CN112287741A (en) | Image processing-based farming operation management method and device | |
| Gan et al. | Fast and accurate detection of lactating sow nursing behavior with CNN-based optical flow and features | |
| CN113762113A (en) | Livestock parturition behavior monitoring method and device | |
| CN111797831A (en) | BIM and artificial intelligence based parallel abnormality detection method for poultry feeding | |
| Sharma | Rise of computer vision and internet of things | |
| Wang et al. | E-YOLO: Recognition of estrus cow based on improved YOLOv8n model | |
| CN115830078B (en) | Multi-target pig tracking and behavior recognition method, computer equipment and storage medium | |
| Kassim et al. | Small object bird detection in infrared drone videos using mask R-CNN deep learning | |
| CN114677614A (en) | Single sow lactation time length calculation method based on computer vision | |
| Korsch et al. | Automated visual monitoring of nocturnal insects with light-based camera traps | |
| Uzun | An intelligent system for detecting Mediterranean fruit fly [Medfly; Ceratitis capitata (Wiedemann)] | |
| BATE | APPLYING DEEP LEARNING TO ESTIMATE FRUIT YIELD IN AGRICULTURE 4.0 SYSTEMS | |
| Stojnic et al. | A Method for Detection of Small Moving Objects in UAV Videos. Remote Sens. 2021, 13, 653 | |
| Mahboob et al. | YOLO v5, v7 and v8: A Performance Comparison for Tobacco Detection in Field | |
| Cruz et al. | Individual Piglet Detection in Weaning Stage using Deep Learning Techniques | |
| CN112700453B (en) | Agricultural water quality supervision method, equipment and medium based on Internet of things | |
| CN114155594A (en) | Behavior recognition method, behavior recognition device, behavior recognition equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |