Himalaya marmot monitoring method based on monitoring video image

Abstract

The embodiment of the invention discloses a Himalayan marmot monitoring method based on a monitoring video image. The method comprises the following steps: acquiring video monitoring of a to-be-monitored area; performing target detection on the video monitoring image to obtain a plurality of marmot targets; inputting each marmot target into a trajectory tracker to obtain a plurality of tracking trajectories; according to speed constraints, non-marmot trajectories which are wrongly recognized are eliminated from the multiple tracking trajectories; according to the time constraint and the distance constraint, the tracking trajectories belonging to the same trajectory in the remaining trajectories are spliced to obtain a plurality of marmot trajectories; calculating the number and the speed of the marmot tracks in different time periods, and analyzing the activity number characteristics and the activity speed characteristics of marmots; and carrying out clustering and contour extraction on the marmot tracks, and analyzing activity range characteristics of marmots. According to the embodiment, the activity characteristics of marmots can be fully detected and mined.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of image processing technology, and in particular to a Himalayan marmot monitoring method based on surveillance video images. Background Art

[0002] The Himalayan marmot is the natural host of plague. By monitoring the Himalayan marmot, we can understand its living environment and activity patterns and prevent the risk of plague.

[0003] In the prior art, patent application CN110443561A discloses a marmot information collection system that integrates multiple modules such as data collection, data editing, and data transmission and analysis. It analyzes the relationship between marmot distribution and vegetation type, vegetation coverage, soil type, and cave type based on location data, attribute data, and image data. Subsequently, during plague monitoring in the natural plague foci of Himalayan marmots in Qinghai Province, detection and data collection were carried out, and combined with relevant marmot ecological data, the distribution characteristics of Himalayan marmots in Qinghai Province at the macro level were inverted, and a remote sensing distribution prediction map was drawn. This method can predict the distribution status of marmots in different regions from a macro perspective, but it is difficult to focus on the changes in the activity characteristics of Himalayan marmots in smaller study areas, and it is difficult to explore the movement characteristics of marmots.

[0004] Patent application CN114841948A discloses a marmot density survey method based on orthophotos from drones. First, drones are used to collect marmot images from different survey areas. These images are then deduplicated and combined into a global image. Finally, intelligent recognition technology is used to count the number, density, and distribution of marmots in the image, thereby determining the number of marmots active in the study area. This approach is difficult to achieve around-the-clock monitoring because drone photography is susceptible to inclement weather and requires on-site personnel for operation or on-call duty. This approach is insufficient in proactive, real-time monitoring of changes in marmot activity characteristics, and there are certain errors in the calculation of marmot density.

[0005] Therefore, it is urgent to propose a Himalayan marmot monitoring method to fully detect and explore the activity characteristics of marmots without human supervision. Summary of the Invention

[0006] The embodiment of the present invention provides a Himalayan marmot monitoring method based on surveillance video images to solve the above technical problems.

[0007] In a first aspect, an embodiment of the present invention provides a Himalayan marmot monitoring method based on surveillance video images, comprising:

[0008] Obtain video surveillance of the area to be monitored;

[0009] Performing target detection on the video surveillance image to obtain a plurality of marmot targets; inputting each marmot target into a trajectory tracker to obtain a plurality of tracking trajectories;

[0010] According to the speed constraint, excluding the misidentified non-marmot tracks from the multiple tracking tracks; according to the time constraint and the distance constraint, splicing the tracking tracks belonging to the same track in the remaining tracks to obtain multiple marmot tracks;

[0011] Calculating the number and speed of the plurality of marmot tracks by time period, and analyzing the number characteristics and speed characteristics of the marmot activity;

[0012] Clustering and contour extraction are performed on the multiple marmot tracks to analyze the characteristics of the marmot's activity range.

[0013] In a second aspect, an embodiment of the present invention provides an electronic device, comprising:

[0014] one or more processors;

[0015] a memory for storing one or more programs,

[0016] When the one or more programs are executed by the one or more processors, the one or more processors implement the Himalayan marmot monitoring method based on surveillance video images described in any embodiment.

[0017] In a third aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the Himalayan marmot monitoring method based on surveillance video images described in any embodiment.

[0018] In summary, the present invention provides a Himalayan marmot monitoring method based on surveillance video images. By using the Himalayan marmot video images of the plague source obtained by monitoring probes, the target detection model and tracking algorithm are used to realize real-time identification and tracking of Himalayan marmots in an unmanned environment, and the marmot activity trajectory is extracted from it from a high temporal resolution perspective, which is then applied to the all-weather monitoring and rapid analysis of the number, speed and range characteristics of Himalayan marmot activities. This overcomes the shortcomings of monitoring the changes in marmot activity characteristics by using drones or manual surveys and statistics in terms of accuracy and real-time performance, and improves the efficiency of investigation and monitoring of natural plague foci of Himalayan marmots. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 is a flow chart of a Himalayan marmot monitoring method based on surveillance video images provided by an embodiment of the present invention;

[0021] Figure 2 is a flowchart of another Himalayan marmot detection method based on surveillance video images provided by an embodiment of the present invention;

[0022] Figure 3 is a schematic diagram of a marmot recognition result in a video image provided by an embodiment of the present invention;

[0023] Figure 4 is a schematic diagram of a marmot tracking result in a video image provided by an embodiment of the present invention;

[0024] Figure 5 is a schematic diagram of the number characteristics of marmot activity provided by an embodiment of the present invention;

[0025] Figure 6 is a schematic diagram of a marmot activity speed characteristic provided by an embodiment of the present invention;

[0026] Figure 7 This is a schematic diagram of the maximum gathering range of marmot activities provided by an embodiment of the present invention;

[0027] Figure 8 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0028] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention are described clearly and completely below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are also within the scope of protection of the present invention.

[0029] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0030] In the description of the present invention, it should also be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood broadly. For example, they may refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on the specific circumstances.

[0031] Figure 1 Flowchart of a Himalayan marmot monitoring method based on monitoring video images provided by an embodiment of the present invention. The method is executed by an electronic device, such as Figure 1 As shown, the specific steps include:

[0032] S110. Construct a marmot image dataset.

[0033] In this embodiment, a video of Himalayan marmots is captured by monitoring probes at the epidemic source to form a video dataset; and a marmot image dataset is created based on the video dataset as a dataset for constructing a marmot detection model.

[0034] In one specific implementation, a rough manual identification process was first used to identify candidate videos containing obvious marmots. OpenCV (Open Source Computer Vision Library) was then used to obtain the video frame rate. The video was then sampled frame by frame at fixed intervals, taking into account the exposure and motion of the target. This yielded a preliminary dataset of Himalayan marmot targets.

[0035] Next, construct an image dataset for deep learning model training. Optionally, use the LabelImg tool to annotate the initial Himalayan marmot dataset. The annotations include the target category and detection box (i.e., bounding box), both of which are stored in a format consistent with the YOLO model. For example, the marmot category name can be defined as "marmot," and the storage format used is as follows:

[0036]

[0037] in, Indicates the id (identification) value of the current category, They represent the x-coordinate value, y-coordinate value of the center point of the normalized detection frame, and the height and width of the detection frame respectively.

[0038] Finally, the image and annotated datasets are divided into training and test sets according to a certain ratio. The training set is used to train a suitable neural network model, enabling it to extract and integrate image target features to detect marmots. The test set is used to verify the model's detection performance.

[0039] S120: Construct a marmot detection model using the marmot image dataset.

[0040] This step builds a marmot detection model based on the above training set and test set through model training, model evaluation, and inference testing.

[0041] In a specific embodiment, the version numbers of the software and hardware used for model training are first determined. In this embodiment, the system environment is Windows 11 version, the CPU (Central Processing Unit) model is a 14-core Intel(R) Core(TM) Ultra 5 125H, the memory size is 32GB, the integrated compilation environment is PyCharmCommunity 2024.3 version, the Python version is 3.12, and the PyTorch version is 2.5.1.

[0042] The YOLO11s model was used for training. Key parameters included image size (imgsz), training epochs (epochs), training mini-batch size (batch), data augmentation strategy (close_mosaic), training device (device), and number of subprocesses (workers). Using these basic parameters, each marmot image was fed into the YOLO11s model, ensuring that the model output consistently approximated the bounding boxes marked in the input image.

[0043] The trained model needs to be evaluated for its effectiveness. Metrics such as mean average precision (mAP), precision, and recall can be used for model evaluation. If the model achieves a high level of accuracy, it can be used as the final marmot detection model (also known as a detector).

[0044] S130, obtaining video surveillance of the area to be monitored; performing target detection on the image of the video surveillance to obtain multiple marmot targets; inputting each marmot target into a trajectory tracker to obtain multiple tracking trajectories.

[0045] The area to be monitored here can be an epidemic source area, which can be the same as or different from the epidemic source in S110. This step combines the detector and the tracker to track the Himalayan marmot in the video image in real time. Optionally, the tracker can use the DeepSORT algorithm. The core tracking parameters involved include the maximum number of survival frames of the trajectory (max_age), the minimum number of initialization frames (n_init), the maximum overlap ratio allowed during non-maximum suppression (nms_max_overlap), the cosine distance threshold for appearance matching (max_cosine_distance), the number of nearest neighbor matches (nn_budget), and the ReID (Re-identification) network (embedder) used to extract target appearance features. Each parameter can be determined according to the target motion state and the complexity of the background in the video image.

[0046] In one specific embodiment, after acquiring video surveillance footage of the area to be monitored, a detector is first used to detect the video image. A confidence threshold is set, and detected targets above this threshold are input into the tracker. The DeepSORT algorithm in the tracker first models the target's motion state using a Kalman filter and constructs a target motion feature matrix using the Mahalanobis distance. The target's appearance state is then modeled using a ReID network and a target appearance feature matrix is constructed using cosine similarity. Finally, the appearance feature matrix and the motion feature matrix are fused to form a comprehensive cost matrix. Based on this matrix, a bipartite graph of the sets of detection boxes in adjacent frames is constructed. The Hungarian algorithm is then used for data association to find the optimal match between the detection boxes, thereby generating a tracking trajectory.

[0047] On the premise of ensuring real-time performance, other methods can also be used to detect and track Himalayan marmots, such as using a variant based on YOLO11 to detect marmots, and using algorithms such as ByteTrack, BotSORT, and StrongSORT to track targets. This embodiment does not impose specific restrictions.

[0048] S140 , excluding misidentified non-marmot tracks from the multiple tracking tracks based on the speed constraint; and splicing the tracking tracks belonging to the same track among the remaining tracks based on the time constraint and the distance constraint to obtain multiple marmot tracks.

[0049] This step screens and optimizes the multiple tracking tracks obtained in S130, excludes misidentified non-marmot tracks, and completes the cut marmot tracks to obtain multiple continuous marmot tracks.

[0050] Alternatively, because the camera is stationary, the Himalayan marmot detection model inherently has fixed errors. Furthermore, due to variations in lighting and background, the YOLO11s model may misidentify similar background objects (such as rocks) as marmots when used for target detection. Therefore, this embodiment uses trajectory velocity constraints to eliminate most misidentified objects. Furthermore, due to short-term occlusions of the target, frequent switching of the same target's trajectory is possible. Therefore, this embodiment uses time and distance constraints to fuse trajectories with different IDs for the same target, improving the accuracy of the trajectory count.

[0051] In one specific embodiment, the average motion velocity of each tracking track is first calculated based on the positions and timestamps of adjacent track points in each tracking track. Specifically, the positions of the track points can be the positions of the center points of the detection frames, and the timestamps of the track points can be the timestamps of the detection frames. The motion velocity between two adjacent detection frames in a single track is calculated by dividing the position difference of the center points by the difference in the timestamps of these two adjacent detection frames. Assuming that there are n detection frames in a single track (n is a natural number greater than 2), n-1 motion velocities can be obtained. The average motion velocity of the single track can be obtained by taking the arithmetic mean of these n-1 motion velocities.

[0052] Then, a maximum threshold and a minimum threshold of the marmot's activity speed are set, and the tracking tracks with an average activity speed lower than the minimum threshold or higher than the maximum threshold are discarded. The maximum threshold and the minimum threshold can be obtained through multiple experiments and manual judgment, and are not specifically limited in this embodiment.

[0053] For the remaining tracking tracks, the following operations can be performed for each two tracks: obtain the position (x1, y1) and time t1 (i.e., timestamp) of the end point of the first track Track1 in the two tracks, and the position (x2, y2) and time t2 of the starting point of the second track Track2. If the time difference between the two positions is within the set duration threshold, and the displacement value of the center point of the detection frame of the two positions in the x direction or the y direction is less than the set distance threshold, then the two tracks can be considered to belong to the same complete track. Finally, all the segmented tracks belonging to the same track are spliced together to form a complete continuous track. For the sake of convenience in distinction and description, this embodiment refers to each spliced continuous track and the track that is not blocked as a marmot track.

[0054] S150, calculating the number and speed of the plurality of marmot tracks by time period, and analyzing the number characteristics and speed characteristics of the marmot's activities; clustering and contour extraction of the plurality of marmot tracks, and analyzing the characteristics of the marmot's activity range.

[0055] This step extracts and analyzes marmot activity characteristics based on their trajectories. Optionally, changes in marmot activity patterns, such as number, speed, and range, can be monitored to further analyze their living environment and activity patterns, helping to prevent plague risks.

[0056] First, let's examine the activity count characteristics of marmots. Optionally, these can be calculated on a daily and hourly scale, representing the number of marmot tracks detected in the video imagery per day and hour, respectively. This metric serves as a key indicator of changes in the activity characteristics of Himalayan marmots. Sudden increases in the number of marmot tracks can be identified as periods of abnormal activity.

[0057] Next, we will introduce the marmot's activity speed characteristics. Optionally, based on the average activity speed obtained in S140, the average activity speed of all marmot trajectories in each time period can be calculated. The periods when the average activity speed suddenly increases or decreases represent periods of abnormal activity speed. The activity speed characteristics are closely related to the behavior patterns of Himalayan marmots and changes in the surrounding environment. When the overall movement speed of the marmot suddenly increases or decreases in a short period of time, it indicates that its movement state is abnormal.

[0058] Finally, we will introduce the characteristics of the marmot's activity range. The activity range characteristics are the areas where the marmot prefers to live in the video image, that is, the spatial characteristics of the marmot's activities. In a specific embodiment, the extraction and analysis of the activity range may include the following steps:

[0059] Step 1: Mask the video image. In video images captured by the same camera, in addition to the area where Himalayan marmots typically move, other irrelevant areas (such as background areas like the sky) can be removed through masking. Alternatively, ArcGIS Pro can be used to construct a planar .shp file of the irrelevant areas, which can then be masked. The remaining area becomes the correct area for analysis.

[0060] Step 2: Cluster the trajectories in the correct area to be analyzed. The marmot activity trajectory data itself is disorganized, so a clustering algorithm can be used to cluster all trajectory point data, and each cluster obtained corresponds to a marmot activity gathering area. Optionally, the DBSCAN algorithm can be used for clustering, which defines a "cluster" as an area with sufficient density and has better performance in non-convex data sets. The algorithm includes two core parameters: the minimum number of included points (min_samples) and the scanning radius (eps). Its basic process can be described as first randomly selecting a point in the area as the starting point, looking for a point smaller than eps, and continuing to look for points smaller than eps based on this new point until no new points are generated. If the number of these points is greater than the min_samples number, then it can be defined as a cluster.

[0061] Step three, extract trajectory contours based on clusters. In this embodiment, concave hull contour extraction is performed on each cluster with the largest number of trajectory points, and at least one clustered area of marmot activity is obtained. Marmot trajectory data is essentially a series of points. After the trajectory clustering operation, a cluster of marmot activity can be obtained. The regional contour of the cluster is uneven, and the commonly used convex hull algorithm is difficult to extract the contour range. Therefore, this embodiment uses the Alpha Shapes algorithm to extract the outer contour of the cluster. The principle of the two-dimensional Alpha Shape algorithm is to draw a circle with a radius of alpha through any two trajectory points among all trajectory points. If the distance between the two points is less than twice alpha, there will be two such circles; if there are no other points in any circle, it can be considered that the line formed by the connection of the two points is one of the boundary lines, thereby gradually obtaining the concave hull contour of the cluster group. This contour can maintain the concave area in the region and obtain a more accurate contour range.

[0062] Step 4: Extraction of activity area features. Taking any gathering area as an example, the texture features, brightness, contrast and edge features inside and outside the gathering area are used to analyze the characteristics of the marmot's activity range. This step extracts image features inside and outside the gathering area, and analyzes the characteristics of the marmot's activity gathering area based on the image features inside and outside the area. Optionally, the main image features involved include texture features, brightness, contrast and edge features. Among them, texture features reflect the visual homogeneity phenomenon in the image, which is manifested as the organized and structured arrangement properties of the object surface. The gray level co-occurrence matrix can be used to obtain the contrast, correlation, energy and homogeneity of the image texture as texture features; image edge features reflect the distribution of high-frequency signals in the image. The Canny algorithm can be used to calculate the gradient distribution in different directions as image edge features; brightness and contrast can be obtained by solving the mean and standard deviation of the grayscale values in different areas. The texture features, brightness, contrast and edge features inside and outside the gathering area can be used to analyze the activity characteristics of marmots. For example, the activity range of marmots in a new area can be predicted based on the texture features inside and outside the gathering area.

[0063] Figure 2 This is a flowchart of another Himalayan marmot detection method based on surveillance video images provided by an embodiment of the present invention. Step 1, Step 2, and Step 3 in the figure represent three main steps respectively, and Track represents the track. The entire process of the above method can also be combined with Figure 2 Understand.

[0064] To demonstrate the monitoring results of the above method, this example applies the method to video observation data from a Himalayan marmot plague focus in Lhasa from August 22 to September 14, 2024. Each video in the video observation data is approximately 30 minutes long. The YOLO11s model and DeepSORT algorithm were used to detect and track Himalayan marmots. Table 1 shows the parameters of the monitoring equipment, the training parameters of the YOLO11s model, and the DeepSORT tracking parameters.

[0065] Table 1 Experimental parameter values

[0066]

[0067] Among them, the YOLO11s model stopped training at the 245th round and tended to converge. The model performed best in the 164th round, with mAP50 of 0.938, mAP50-95 of 0.470, Precision of 0.947, and Recall of 0.882. Among them, mAP50 is the average of the average precision calculated under the condition of an IoU threshold of 0.5, and mAP50-95 is the average of the average precision calculated under the condition of an IoU threshold of [0.5, 0.95]. IoU represents the intersection-over-union ratio between the predicted detection box and the actual bounding box. At the same time, the changes in the loss of the model training set and the test set are relatively coordinated, which proves that the marmot detection model has a certain degree of generalization. Figure 3 This is a schematic diagram of the marmot recognition results in the video image. The values in the figure are the confidence values of the marmot detection box marmot.

[0068] Based on the marmot detection model, 0.7 is used as the confidence threshold of the detector. Targets with a value greater than this threshold are input into the tracker. The DeepSORT algorithm is used for target tracking. The processing rate can reach 19fps (frames per second), which is higher than the video frame rate and can be considered as real-time processing. Figure 4 Figure 2 shows the results of tracking marmots in video images. The above operations were performed on all surveillance videos, and 765 tracking tracks were finally extracted from 602 videos.

[0069] Then, based on the activity characteristics of marmots in the image, the minimum threshold (min_v) and maximum threshold (max_v) of the marmot activity speed were set to 2 pixels per second and 100 pixels per second, respectively. The time constraint was 60 seconds and the distance constraint was 50 pixels. Finally, 121 marmot tracks were screened and fused from the 765 tracking tracks.

[0070] Based on 121 marmot tracks, we can further monitor the number of marmots' activities, activity speed, and activity range characteristics, and analyze the marmot's activity characteristics. The final analysis results are: the Himalayan marmot's overall movement is slow, there is a certain degree of clustering in the activity range, the daytime activity frequency is generally low and locally high, and the hourly activity frequency has clustered peaks and stable valleys. In the activity number characteristics, frequent marmot activities were detected only on 6 days out of 24 days, of which 10 a.m. every day was the peak state of marmot activity, and 19 p.m. was the lowest period; in the activity speed characteristics, the marmot in the image moved an average of 14.42 pixels per second, and the average movement speed was less than 10 pixels in most time periods, indicating that the marmot was in a state of eating at this time. Figure 6The figure represents the change in marmot activity speed. For the activity range feature, we set the eps value in the DBSCAN clustering to 15, the min_samples value to 10, and the alpha value in the Alpha Shape algorithm to 0.03. We found that marmots prefer to move in low-contrast, low-brightness areas that are primarily grassy, with some obstructions such as mounds, rocks, and small bushes. Figure 5 、 Figure 6 They respectively reflect the changing characteristics of marmot activity quantity and activity speed. Figure 7 It represents the largest gathering area of marmot activity (within the red line area). Table 2 reflects the texture features, edge features, and contrast and brightness calculation results of the images within and outside the gathering area.

[0071] Table 2 Experimental results

[0072]

[0073] In summary, this embodiment takes video observation data from a certain epidemic source as an example, and uses a target detection and tracking method based on YOLO11s and DeepSORT to realize the detection and real-time tracking of Himalayan marmot objects in video images, and performs feature extraction based on the marmot trajectory, monitoring the changes in the marmot's movement status from three aspects: quantity, speed, and activity range. The solution is feasible and the conclusion is reliable.

[0074] In summary, the present invention provides a Himalayan marmot monitoring method based on surveillance video images to improve the survey results and efficiency of plague natural foci. First, the video images are processed using the YOLO11 model and the DeepSORT algorithm to detect and track marmot targets in real time; then, the marmot activity characteristics in natural situations are mined and quantified based on the marmot activity trajectory, and changes in its activity state are monitored. This method can make up for the shortcomings of using drones or traditional manual surveys and statistics to monitor changes in marmot activity characteristics in terms of initiative and real-time performance, and realize all-weather monitoring and rapid analysis of Himalayan marmot activity characteristics in an unmanned environment. At the same time, this embodiment tracks the trajectory of targets with a confidence level greater than a set threshold, and automatically fills in some missed trajectory points through trajectory tracking, making up for some missing marmot positions in target detection, and providing more trajectory point information for the clustering algorithm.

[0075] More specifically, the use of target detection and trajectory tracking to extract the marmot's activity range has the following advantages: if only target detection is performed on each frame of the image without trajectory tracking, the target may be missed. For example, if the target is occluded for a short time (obstructed by grass, stones, etc.) or the motion state changes, it will be difficult to detect the target in the image using the detector; while the "target detection + trajectory tracking" method can track the same target. Even if the target is occluded for a short time, the range of the target in the image can be roughly determined by difference, so that a more accurate marmot activity range can be obtained.

[0076] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention is shown in FIG. Figure 8 As shown, the device includes a processor 60, a memory 61, an input device 62 and an output device 63; the number of processors 60 in the device can be one or more. Figure 8 In the embodiment, a processor 60 is used as an example; the processor 60, the memory 61, the input device 62 and the output device 63 in the device can be connected by a bus or other means. Figure 8 The bus connection is taken as an example.

[0077] Memory 61, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the Himalayan marmot monitoring method based on surveillance video images in the embodiments of the present invention. Processor 60 executes the software programs, instructions, and modules stored in memory 61 to perform various functional applications and data processing of the device, thereby implementing the aforementioned Himalayan marmot monitoring method based on surveillance video images.

[0078] The memory 61 may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function; the data storage area may store data generated based on the use of the terminal. Furthermore, the memory 61 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state memory device. In some instances, the memory 61 may further include memory remotely located relative to the processor 60, and these remote memories may be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0079] The input device 62 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 63 may include a display device such as a display screen.

[0080] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the Himalayan marmot monitoring method based on surveillance video images of any embodiment is implemented.

[0081] The computer storage medium of the embodiments of the present invention may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device.

[0082] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0083] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0084] Computer program code for performing the operations of the present invention can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as C or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A Himalayan marmot monitoring method based on surveillance video images, characterized in that: include: Obtain video surveillance of the area to be monitored; Performing target detection on the video surveillance image to obtain multiple marmot targets; Each marmot target is input into the trajectory tracker to obtain multiple tracking trajectories; According to the speed constraint, excluding the misidentified non-marmot tracks from the multiple tracking tracks; and according to the time constraint and the distance constraint, splicing the tracking tracks belonging to the same track in the remaining tracks to obtain multiple marmot tracks; Calculating the number and speed of the plurality of marmot tracks by time period, and analyzing the number characteristics and speed characteristics of the marmot activity; Clustering and contour extraction are performed on the multiple marmot tracks to analyze the characteristics of the marmot's activity range.

2. The method according to claim 1, wherein the step of performing target detection on the video surveillance image to obtain a plurality of marmot targets comprises: The YOLO model is trained using the labeled marmot images so that the trained model can display the marmots in the image as detection boxes. Using the trained YOLO model, perform target detection on the video surveillance image to obtain multiple detection frames and their confidence levels; Multiple detection boxes with confidence levels higher than a set threshold are determined to be multiple marmot targets.

3. The method according to claim 1, characterized in that The step of excluding misidentified non-marmot tracks from the plurality of tracking tracks according to the speed constraint includes: According to the activity of marmots, the highest and lowest thresholds of marmot activity speed are determined; Calculate the average activity speed of each tracking track based on the positions and timestamps of adjacent track points in each tracking track; Tracking tracks with an average activity speed higher than the maximum threshold or lower than the minimum threshold are determined as misidentified non-marmot tracks and are excluded.

4. The method according to claim 1, wherein According to the time constraint and the distance constraint, the tracking tracks belonging to the same track in the remaining tracks are spliced together to obtain multiple marmot tracks, including: Traverse each remaining trajectory; If the time stamp difference between the end point of one remaining track and the starting point of another remaining track is less than a set time threshold, and the displacement value between the end point and the starting point along the x-direction or y-direction is less than a set distance threshold, the remaining track and the other remaining track are determined to belong to the same tracking track; All tracking tracks belonging to the same track are spliced into one marmot track.

5. The method according to claim 1, wherein The calculating the number and speed of the plurality of marmot tracks by time period and analyzing the number characteristics and speed characteristics of the marmot activity includes: The number of marmot tracks in each period was counted, and periods with sudden increases in numbers were identified as periods of abnormal activity. The average activity speed of the marmot's trajectory in each period was calculated, and the periods with sudden increases or decreases in the average activity speed were identified as periods of abnormal activity speed. Pay attention to environmental changes during abnormal periods to analyze the factors affecting marmot activities.

6. The method according to claim 1, characterized in that The clustering and contour extraction of the plurality of marmot tracks and analysis of the activity range characteristics of the marmot include: Clustering the track points in the plurality of marmot tracks to obtain a plurality of clusters; The concave hull contour of the clusters with the largest number of trajectory points is extracted to obtain the gathering area of marmot activities; The texture features, brightness, contrast and edge features inside and outside the aggregation area are extracted respectively for analyzing the activity range characteristics of the marmot.

7. The method according to claim 6, characterized in that After extracting the texture features, brightness, contrast and edge features inside and outside the clustering area respectively, the method further includes: The activity range of marmots in the new area is predicted based on the texture features inside and outside the aggregation area.

8. The method according to claim 6, characterized in that The concave hull contour extraction of the cluster with the largest number of trajectory points includes: The Alpha Shapes algorithm is used to extract the concave hull contours of the clusters with the largest number of trajectory points.

9. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the Himalayan marmot monitoring method based on surveillance video images as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the program is executed by a processor, the method for monitoring Himalayan marmots based on surveillance video images as described in any one of claims 1 to 8 is implemented.

Images