Unmanned aerial vehicle target intelligent identification and perception system and method based on deep learning

Abstract

The invention discloses an unmanned aerial vehicle target intelligent identification and perception system and method based on deep learning, and relates to the technical field of artificial intelligence, and the method comprises the steps: extracting the importance feature of each target region, calculating a region importance score, generating a target region monitoring plan, and obtaining continuous frame images; an improved YOLOv5 model is adopted, continuous frames of images are input, a bounding box of a suspected target wild animal is output and mapped to an actual geographic coordinate system, and actual position coordinates are obtained; key points of the t-th frame image are detected, coordinates of the key points are extracted, visibility scores and confidence coefficients are calculated, and feature descriptors are formed; performing target detection and key point detection on the (t + 1) th frame of image to obtain actual position coordinates and feature descriptors of candidate targets, performing track matching, updating the actual position coordinates and the feature descriptors of the suspected target wild animals if the actual position coordinates and the feature descriptors are matched, and defining new target animals if the actual position coordinates and the feature descriptors are not matched; counting according to the track matching result to obtain the total number of the target wild animals.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a system and method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning. Background Art

[0002] UAV target intelligent recognition and perception is an important research field in UAV systems. It mainly involves the use of computer vision, image processing, pattern recognition and other technologies to achieve automatic detection, recognition, tracking and other functions of UAVs.

[0003] The Chinese patent with authorization announcement number CN116310894B discloses a small-sample, small-target Tibetan antelope intelligent identification method based on drone remote sensing. The method collects drone orthophotos through drones; constructs auxiliary sets and a small sample library of Tibetan antelopes from drones; constructs a small-sample deep learning model with context-aware fusion and comparative analysis; determines the model loss function; and uses the auxiliary sets and a small sample library of Tibetan antelopes from drones to train the small-sample deep learning model with context-aware fusion and comparative analysis.

[0004] When using drones to identify wild animals and count their numbers, wild animal counting statistics are prone to duplication or omissions due to the small differences in appearance between individual wild animals and the complexity of animal activities. Summary of the invention

[0005] The present application aims to solve at least one of the technical problems in the related art to a certain extent. To this end, one purpose of the present application is to propose a drone target intelligent recognition perception system and method based on deep learning to improve the accuracy of wildlife population statistics in the target area.

[0006] One aspect of the present application provides a method for intelligent recognition and perception of drone targets based on deep learning, including: Step S100: Divide the target area according to the geographical scope of the wildlife reserve, extract the importance features of each target area and calculate the regional importance score, generate a target area monitoring plan, and the drone acquires continuous frame images based on the target area monitoring plan; Step S200: using the improved YOLOv5 model to perform target detection, inputting continuous frame images, outputting the bounding box of the suspected target wild animal in each frame image and mapping it to the actual geographic coordinate system, and obtaining the actual position coordinates of the suspected target wild animal in each frame image; Step S300: perform key point detection on the t-th frame image, extract the key point coordinates of the suspected target wild animal, calculate the visibility score of the key point and the confidence of each key point, and the K key points with the highest confidence constitute the feature descriptor; Step S400: Perform target detection and key point detection on the t+1th frame image to obtain the actual position coordinates and feature descriptors of the candidate target, and perform trajectory matching between the suspected target wild animal and the candidate target. If they match, the actual position coordinates and feature descriptors of the suspected target wild animal are updated; if they do not match, the candidate target is defined as a new target animal; Step S500: According to the trajectory matching results of the suspected target wild animals and the candidate targets, the identified suspected target wild animals and new target animals are counted by a counter to obtain the total number of target wild animals in the target area.

[0007] Specifically, the target area is divided according to the geographical scope of the wildlife reserve, the importance features of each target area are extracted and the regional importance score is calculated, and a target area monitoring plan is generated. The specific method for the drone to obtain continuous frame images based on the target area monitoring plan is: Step S110: Divide the geographical scope of the wildlife reserve into a target areas and assign unique numbers, record basic information and historical monitoring data of each target area, and establish a target area database, wherein the basic information includes food abundance, shelter conditions and water availability of the target area, and the area of ​​the target area, and the historical monitoring data includes the frequency of occurrence of target wildlife, the population density of target wildlife, the frequency of each behavior pattern of target wildlife being observed, the time of each observation of target wildlife activity, the nearest neighbor distance between the target wildlife and its adjacent wildlife, and the number of target wildlife; Step S120: extracting the importance features of the target area from the target area database, wherein the importance features include: the frequency of occurrence of target wild animals; , Population density of target wildlife and behavioral pattern diversity index , Regional Habitat Quality Index , Time Distribution Regularity Index , spatial distribution law index ; Step S130: Calculate the importance score of the target area based on the importance feature ; Step S140: normalizing the importance scores of the target areas, and mapping the importance scores of the target areas to various monitoring parameters according to the preset mapping functions of the various monitoring parameters; the monitoring parameters include: monitoring frequency MF, single monitoring duration MD, and number of drones DN; Step S150: generating a target area monitoring plan according to the monitoring parameters of each target area, and the UAV acquires continuous frame images according to the target area monitoring plan.

[0008] Specifically, the improved YOLOv5 model is used for target detection, continuous frame images are input, the bounding box of the suspected target wild animal in each frame image is output and mapped to the actual geographic coordinate system, and the specific method for obtaining the actual position coordinates of the suspected target wild animal in each frame image is: Step S210: Use the improved YOLOv5 model to detect wildlife targets and set the confidence threshold , take continuous frame images as input, perform feature scaling on each input frame image to obtain a feature pyramid, input the feature pyramid into the improved YOLOv5 model, extract the target confidence prediction value, bounding box coordinate prediction value and category probability prediction value corresponding to the prediction bounding box with a confidence greater than the confidence threshold at each size, remove redundant prediction bounding boxes, and convert the target confidence prediction value of the remaining prediction bounding boxes into , bounding box coordinate prediction value and class probability prediction value Output, where Indicates the horizontal offset of the center of the bounding box relative to the upper left corner of the predicted bounding box. Indicates the vertical offset of the center of the bounding box relative to the upper left corner of the predicted bounding box. is the center point of the bounding box, are the bounding box width and the bounding box height, represents the probability that the predicted bounding box belongs to the nth target wildlife category, where n is the total number of target wildlife categories, and the remaining predicted bounding boxes are used as bounding boxes of suspected target wildlife; The specific construction method of the improved YOLOv5 model is: Step S211: construct an improved YOLOv5 model, use CSPDarknet53 as the backbone network, add an attention mechanism at each stage of the backbone network, the attention mechanism consists of two fully connected layers and a Sigmoid activation function, and assume that the input feature map is X and the number of channels is C; Step S212: The calculation process of the attention mechanism is as follows: perform global average pooling on each channel of X to obtain a feature vector Z with a size of 1×1×C; pass Z through the first fully connected layer FC1 to obtain a feature vector with a size of 1×1× The feature vector U of , where r is the scaling factor, is used to control the parameter amount of the attention mechanism; U is passed through the second fully connected layer FC2 to obtain a feature vector V of size 1×1×C; each element in V is activated by the Sigmoid activation function to obtain the attention weight vector A; the input feature map X is multiplied by the attention weight vector A to obtain the modulated feature map ; Step S213: using the FPN structure to fuse feature maps of different sizes, and constructing a feature pyramid through upsampling and splicing operations; Step S214: setting a detection head at each size of the feature pyramid, wherein the detection head is composed of a convolutional layer and a fully connected layer, and the detection head outputs target confidence, bounding box coordinates, and category probability; Step S215: using a binary cross entropy loss function to calculate the confidence loss between the target confidence prediction value and the target confidence actual value, using a GIOU loss function to calculate the bounding box loss between the bounding box coordinate prediction value and the bounding box coordinate actual value, using a multivariate cross entropy loss function to calculate the category loss between the category probability prediction value and the category probability actual value, and weightedly summing the confidence loss, bounding box loss, and category loss to obtain the total loss function of the improved YOLOv5 model; Step S216: Obtain wildlife images of historical time, annotate each wildlife image with the category label and bounding box coordinates of the target wildlife, perform data enhancement on the wildlife images to obtain training images, use the annotated training images as input data, output corresponding target confidence prediction values, bounding box coordinate prediction values, and category probability prediction values ​​based on the improved YOLOv5 model, and use the value that minimizes the total loss function as the training goal. When the total loss function converges, the training is completed, and a trained improved YOLOv5 model is obtained.

[0009] Step S220: According to the class probability prediction values ​​of the suspected target wild animals, the target wild animal class corresponding to the largest class probability prediction value is selected as the preliminary recognition result; Step S230: According to the size of the feature map , calculate the target center coordinates on the feature map , map the target center coordinates on the feature map to the continuous frame images, and calculate the normalized coordinates of the target center coordinates in the continuous frame images ; Step S240: Based on the normalized coordinates and the current GPS coordinates of the drone , UAV image resolution , map the bounding box of the suspected target wild animal to the actual geographic coordinate system, and obtain the actual position coordinates of the suspected target wild animal in each frame of the continuous frame image .

[0010] Specifically, the method of performing key point detection on the t-th frame image, extracting the key point coordinates of the suspected target wild animal, calculating the visibility score of the key point and the confidence of each key point, and forming the feature descriptor with the K key points with the highest confidence is as follows: Step S310: Extract the key point coordinates of the suspected target wild animal in the t-th frame image based on the key point detection technology , the visibility score of the key point is calculated based on the conditional probability of the key point visibility ; Step S320: Set the visibility score threshold, and calculate the confidence of each key point based on the key point coordinates and visibility score. , when the visibility score is greater than or equal to the visibility score threshold, the confidence of the key point is equal to its visibility score, otherwise the confidence of the key point is equal to 0; Step S330: Select the K key points with the highest confidence as the key point features of the suspected target wild animal, forming the feature descriptor of the suspected target wild animal in the t-th frame image. .

[0011] Specifically, the target detection and key point detection are performed on the t+1 frame image to obtain the actual position coordinates and feature descriptors of the candidate target, and the trajectory of the suspected target wild animal is matched with the candidate target. If they match, the actual position coordinates and feature descriptors of the suspected target wild animal are updated; if they do not match, the candidate target is defined as a new target animal. The specific method is: Step S410: performing target detection on the t+1th frame image based on the improved YOLOv5 model, outputting the target confidence prediction value, bounding box coordinate prediction value and category probability prediction value of the candidate target in the t+1th frame image, and calculating the actual position coordinates of the candidate target; Step S420: Perform key point detection on the t+1th frame image to obtain the feature descriptor of the candidate target in the t+1th frame image ; Step S430: Calculate the similarity between the candidate target and the suspected target wildlife based on the feature descriptor of the candidate target and the feature descriptor of the suspected target wildlife. ; Step S440: Preset similarity threshold If the similarity between the candidate target and the suspected target wild animal is greater than the similarity threshold, the candidate target and the suspected target wild animal are considered to be matched successfully, and the actual position coordinates of the suspected target wild animal are updated to the actual position coordinates of the candidate target, and the feature descriptor of the suspected target wild animal is updated to the feature descriptor of the candidate target; Step S450: If the similarity between the candidate target and the suspected target wild animal is less than or equal to the similarity threshold, the suspected target wild animal is considered to be lost in the t+1th frame image, and the candidate target in the t+1th frame image is defined as a new target animal.

[0012] Specifically, according to the track matching results of the suspected target wild animals and the candidate targets, the identified suspected target wild animals and new target animals are counted by a counter to obtain the total number of target wild animals in the target area. Step S510: setting a counter count with an initial value of 0 for counting the total number of target wild animals in the target area; Step S520: traverse all the suspected target wild animals that have been successfully matched, and determine whether the suspected target wild animals have been counted. If the suspected target wild animals have not been counted, add 1 to the value of the counter count; otherwise, the value of the counter count remains unchanged; Step S530: traverse all new target animals to determine whether the new target animal has been counted. If the new target animal has not been counted, add 1 to the value of the counter count; otherwise, the value of the counter count remains unchanged. Step S540: The final value of the counter count is used as the total number of target wild animals in the target area.

[0013] One aspect of the present application provides a UAV target intelligent recognition and perception system based on deep learning, comprising: The image monitoring acquisition module is used to divide the target area according to the geographical scope of the wildlife reserve, extract the importance features of each target area and calculate the regional importance score, generate a target area monitoring plan, and the drone acquires continuous frame images based on the target area monitoring plan; The image target detection module is used to perform target detection using the improved YOLOv5 model, input continuous frame images, output the bounding box of the suspected target wild animal in each frame image and map it to the actual geographic coordinate system to obtain the actual position coordinates of the suspected target wild animal in each frame image; The image key point detection module is used to detect key points of the t-th frame image, extract the key point coordinates of the suspected target wild animal, calculate the visibility score of the key point and the confidence of each key point, and the K key points with the highest confidence constitute the feature descriptor; The animal track matching module is used to perform target detection and key point detection on the t+1 frame image to obtain the actual position coordinates and feature descriptors of the candidate target, and to match the track of the suspected target wild animal with the candidate target. If they match, the actual position coordinates and feature descriptors of the suspected target wild animal are updated; if they do not match, the candidate target is defined as a new target animal; The animal number counting module is used to count the suspected target wild animals and new target animals identified through a counter based on the trajectory matching results of the suspected target wild animals and the candidate targets, so as to obtain the total number of target wild animals in the target area.

[0014] One aspect of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning are implemented.

[0015] One aspect of the present application provides a readable storage medium, which stores a computer program, and the computer program is suitable for loading by a processor to execute the steps in a method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning.

[0016] Compared with the existing technology, the UAV target intelligent recognition and perception system and method based on deep learning proposed in this application have the following advantages: This application proposes a number of novel characteristic indicators, such as behavioral pattern diversity index, regional habitat quality index, spatiotemporal distribution law index, etc., designs a complex regional importance scoring formula, a nonlinear comprehensive multi-indicator, comprehensively reflects the regional ecological importance, and establishes a mapping function between scoring and monitoring parameters to make the scoring results more operational.

[0017] This application introduces attention mechanism and multi-scale feature fusion based on the characteristics of wild animal targets, improves detection accuracy, establishes a mapping relationship between image coordinates and geographic coordinates, can solve the actual position of the target in real time, realizes real-time positioning of the target, overcomes the limitation of traditional methods that require prior modeling, and has better real-time and environmental adaptability.

[0018] This application quantifies the quality of key points by calculating the key point visibility score, selects high-quality key points according to the confidence level to construct feature descriptors, obtains representative target feature descriptions, and improves the feature robustness of target matching and recognition.

[0019] This application achieves cross-frame tracking of wild animal targets by calculating similarity based on key point feature descriptors and performing trajectory matching, effectively improving the performance of recognition and tracking.

[0020] This application introduces a counter to avoid repeated counting. Based on the results of target detection, key point detection, and trajectory matching, it provides a clear target quantity counting method, thereby improving the accuracy of wild animal quantity counting in the target area. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A method flow chart of the deep learning-based intelligent recognition and perception method for unmanned aerial vehicle targets provided in this application; Figure 2 Flowchart of the target detection method provided for this application; Figure 3 Flow chart of the key point detection method provided for this application; Figure 4 Functional module diagram of the deep learning-based UAV target intelligent recognition and perception system provided for this application. DETAILED DESCRIPTION

[0022] In order to better understand the present application, various aspects of the present application will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present application and do not limit the scope of the present application in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0023] In the accompanying drawings, the size, dimensions and shapes of the elements have been slightly adjusted for ease of illustration. The drawings are for illustration only and are not strictly drawn to scale. As used herein, the terms "substantially", "approximately" and similar terms are used as terms of approximation, not as terms of degree, and are intended to illustrate the inherent deviations in measurements or calculations that will be recognized by those of ordinary skill in the art. In addition, in this application, the order in which the steps are described does not necessarily represent the order in which these processes occur in actual operation, unless otherwise specified or can be derived from the context.

[0024] It should also be understood that expressions such as "including", "comprising", "having", "containing" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present application, "may" is used to mean "one or more embodiments of the present application". And, the term "exemplary" is intended to refer to an example or illustration.

[0025] Unless otherwise defined, all words (including engineering terms and scientific and technological terms) used in this article have the same meaning as those commonly understood by ordinary technicians in the field to which this application belongs. It should also be understood that unless there is a clear explanation in this application, words defined in commonly used dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0026] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0027] Example 1 like Figure 1 As shown, the UAV target intelligent recognition and perception method based on deep learning provided by this application includes: Step S100: Divide the target area according to the geographical scope of the wildlife reserve, extract the importance features of each target area and calculate the regional importance score, generate a target area monitoring plan, and the drone acquires continuous frame images based on the target area monitoring plan; The target area is divided according to the geographical scope of the wildlife reserve, the importance features of each target area are extracted and the regional importance score is calculated, and a target area monitoring plan is generated. The specific method for the drone to obtain continuous frame images based on the target area monitoring plan is as follows: Step S110: Divide the geographical scope of the wildlife reserve into a target areas and assign unique numbers, record basic information and historical monitoring data of each target area, and establish a target area database, wherein the basic information includes food abundance, shelter conditions and water availability of the target area, and the area of ​​the target area, and the historical monitoring data includes the frequency of occurrence of target wildlife, the population density of target wildlife, the frequency of each behavior pattern of target wildlife being observed, the time of each observation of target wildlife activity, the nearest neighbor distance between the target wildlife and its adjacent wildlife, and the number of target wildlife; Step S120: extracting the importance features of the target area from the target area database, wherein the importance features include: the frequency of occurrence of target wild animals; , Population density of target wildlife and behavioral pattern diversity index , Regional Habitat Quality Index , Time Distribution Regularity Index , spatial distribution law index ; The behavioral pattern diversity index reflects the richness of behavioral patterns exhibited by target wildlife in the target area; The calculation formula of the behavior pattern diversity index is: ,in, is the frequency at which the jth behavior pattern is observed; The regional habitat quality index reflects habitat factors such as food abundance, shelter conditions, and water availability in the target area; The calculation formula of the regional habitat quality index is: ,in, Indicates the abundance of food. Indicates the comfort level of concealed conditions, Indicates water availability, They represent the weight coefficients of food abundance, shelter comfort, and water availability, respectively; The weight coefficients of food abundance, shelter comfort, and water source availability are set by those skilled in the art based on experience, wherein food abundance, shelter comfort, and water source availability are determined by those skilled in the art through professional evaluation based on food distribution, shelter conditions, and water source distribution in the target area; The temporal distribution regularity index reflects the temporal regularity and stability of the activities of target wild animals; The calculation formula of the time distribution regularity index is: , where NI is the number of drone observations, ni is the serial number of drone observations, ni∈{1,2,...,NI}, is the time angle at which the target wildlife activity is observed for the nith time; The time angle at which the target wildlife activity is observed 0~360 degrees is used to represent 24 hours a day; The spatial distribution regularity index reflects the spatial distribution pattern and aggregation degree of target wildlife in the region. The value range of NNI is (0, 2.15). A value close to 1 indicates random distribution, a value less than 1 indicates aggregated distribution, and a value greater than 1 indicates uniform distribution. The calculation formula of the spatial distribution law index is: ,in, is the nearest neighbor distance between the nath target wild animal and its adjacent wild animals, refers to the arithmetic mean of the nearest neighbor distances, The target area. Refers to the target wildlife population; Step S130: Calculate the importance score of the target area based on the importance feature ; The calculation formula of the target area importance score is: ,in, , , , , , The frequency of occurrence of target wild animals , Population density of target wildlife and behavioral pattern diversity index , Regional Habitat Quality Index , Time Distribution Regularity Index , spatial distribution law index The weight coefficient of The frequency of occurrence of the target wildlife , Population density of target wildlife and behavioral pattern diversity index , Regional Habitat Quality Index , Time Distribution Regularity Index , spatial distribution law index The weight coefficient is set by those skilled in the art based on experience.

[0028] Step S140: normalizing the importance scores of the target areas, and mapping the importance scores of the target areas to various monitoring parameters according to the preset mapping functions of the various monitoring parameters; the monitoring parameters include: monitoring frequency MF, single monitoring duration MD, and number of drones DN; The mapping function of the monitoring frequency is: ,in, is the maximum value of monitoring frequency, Indicates rounding to the nearest integer; Preferably, 30 times per month; The mapping function of the single monitoring duration is: ,in, The maximum value of a single monitoring duration; Preferably, 2 hours; The mapping function of the number of drones is: ,in, is the maximum number of drones, Indicates rounding up; Step S150: generating a target area monitoring plan according to the monitoring parameters of each target area, and the UAV acquires continuous frame images according to the target area monitoring plan; The above steps designed a complex regional importance scoring formula, which comprehensively considers the nonlinear contribution and interactive effects of multiple characteristic indicators. This scoring method can more comprehensively and accurately reflect the ecological importance of each target area and provide more targeted guidance for drone monitoring.

[0029] Step S200: using the improved YOLOv5 model to perform target detection, inputting continuous frame images, outputting the bounding box of the suspected target wild animal in each frame image and mapping it to the actual geographic coordinate system, and obtaining the actual position coordinates of the suspected target wild animal in each frame image; The improved YOLOv5 model is used for target detection, continuous frame images are input, the bounding box of the suspected target wild animal in each frame image is output and mapped to the actual geographic coordinate system, and the specific method for obtaining the actual position coordinates of the suspected target wild animal in each frame image is: Step S210: Use the improved YOLOv5 model to detect wildlife targets and set the confidence threshold , take continuous frame images as input, perform feature scaling on each input frame image to obtain a feature pyramid, input the feature pyramid into the improved YOLOv5 model, extract the target confidence prediction value, bounding box coordinate prediction value and category probability prediction value corresponding to the prediction bounding box with a confidence greater than the confidence threshold at each size, remove redundant prediction bounding boxes, and convert the target confidence prediction value of the remaining prediction bounding boxes into , bounding box coordinate prediction value and class probability prediction value Output, where Indicates the horizontal offset of the center of the bounding box relative to the upper left corner of the predicted bounding box. Indicates the vertical offset of the center of the bounding box relative to the upper left corner of the predicted bounding box. is the center point of the bounding box, are the bounding box width and the bounding box height, represents the probability that the predicted bounding box belongs to the nth target wildlife category, where n is the total number of target wildlife categories, and the remaining predicted bounding boxes are used as bounding boxes of suspected target wildlife; The predicted bounding box coordinates include the center point of the bounding box, the width of the bounding box, and the height of the bounding box, with the cell of each predicted bounding box as the coordinate system; The category probability prediction value refers to the probability of belonging to each target wildlife category; The value of the confidence threshold is set by those skilled in the art based on experience.

[0030] The specific construction method of the improved YOLOv5 model is: Step S211: construct an improved YOLOv5 model, use CSPDarknet53 as the backbone network, add an attention mechanism at each stage of the backbone network, the attention mechanism consists of two fully connected layers and a Sigmoid activation function, and assume that the input feature map is X and the number of channels is C; Step S212: The calculation process of the attention mechanism is as follows: perform global average pooling on each channel of X to obtain a feature vector Z with a size of 1×1×C; pass Z through the first fully connected layer FC1 to obtain a feature vector with a size of 1×1× The feature vector U of , where r is the scaling factor, is used to control the parameter amount of the attention mechanism; U is passed through the second fully connected layer FC2 to obtain a feature vector V of size 1×1×C; each element in V is activated by the Sigmoid activation function to obtain the attention weight vector A; the input feature map X is multiplied by the attention weight vector A to obtain the modulated feature map ; The calculation formula of the attention mechanism is: , , , , ,in, represents global average pooling, It means that Z passes through the first fully connected layer FC1, It means passing U through the second fully connected layer. , are the weight matrices of the first and second fully connected layers, respectively. , are the bias items of the first and second fully connected layers respectively, is the Sigmoid activation function, is channel-by-channel multiplication; Step S213: using the FPN structure to fuse feature maps of different sizes, and constructing a feature pyramid through upsampling and splicing operations; Step S214: setting a detection head at each size of the feature pyramid, wherein the detection head is composed of a convolutional layer and a fully connected layer, and the detection head outputs target confidence, bounding box coordinates, and category probability; Step S215: using a binary cross entropy loss function to calculate the confidence loss between the target confidence prediction value and the target confidence actual value, using a GIOU loss function to calculate the bounding box loss between the bounding box coordinate prediction value and the bounding box coordinate actual value, using a multivariate cross entropy loss function to calculate the category loss between the category probability prediction value and the category probability actual value, and weightedly summing the confidence loss, bounding box loss, and category loss to obtain the total loss function of the improved YOLOv5 model; Step S216: Obtain wildlife images of historical time, annotate each wildlife image with the category label and bounding box coordinates of the target wildlife, perform data enhancement on the wildlife images to obtain training images, use the annotated training images as input data, output corresponding target confidence prediction values, bounding box coordinate prediction values, and category probability prediction values ​​based on the improved YOLOv5 model, and use the value that minimizes the total loss function as the training goal. When the total loss function converges, the training is completed, and a trained improved YOLOv5 model is obtained.

[0031] Step S220: According to the class probability prediction values ​​of the suspected target wild animals, the target wild animal class corresponding to the largest class probability prediction value is selected as the preliminary recognition result; Step S230: According to the size of the feature map , calculate the target center coordinates on the feature map , map the target center coordinates on the feature map to the continuous frame images, and calculate the normalized coordinates of the target center coordinates in the continuous frame images ; The calculation formula of the target center coordinates on the feature map is: , ; The mapping formula for mapping the target center coordinates on the feature map to the continuous frame images is: , ,in, is the target center coordinate mapped to the continuous frame image, is the width of continuous frame images, is the height of continuous frame images; The calculation formula of the normalized coordinates of the target center coordinates in the continuous frame images is: , ; The target center coordinates are based on the global coordinate system of the feature map; Step S240: Based on the normalized coordinates and the current GPS coordinates of the drone , UAV image resolution , map the bounding box of the suspected target wild animal to the actual geographic coordinate system, and obtain the actual position coordinates of the suspected target wild animal in each frame of the continuous frame image ; The mapping formula of the actual position coordinates of the suspected target wild animal is: , ; The above steps optimize and improve the YOLOv5 algorithm based on the characteristics of wild animal targets, introduce the attention mechanism and multi-scale feature fusion strategy, and improve the detection accuracy of wild animals. The above steps achieve real-time calculation of the target position by establishing a mapping relationship between the image coordinate system and the geographic coordinate system, overcoming the limitation of the traditional method that requires the establishment of a complex environmental model in advance, and have better real-time and adaptability.

[0032] Step S300: perform key point detection on the t-th frame image, extract the key point coordinates of the suspected target wild animal, calculate the visibility score of the key point and the confidence of each key point, and the K key points with the highest confidence constitute the feature descriptor; The specific method of performing key point detection on the t-th frame image, extracting the key point coordinates of the suspected target wild animal, calculating the visibility score of the key point and the confidence of each key point, and forming the feature descriptor with the K key points with the highest confidence is as follows: Step S310: Extract the key point coordinates of the suspected target wild animal in the t-th frame image based on the key point detection technology , the visibility score of the key point is calculated based on the conditional probability of the key point visibility ; The calculation formula of the visibility score of the key point is: ,in, is the body part label to which the i-th key point belongs, The category of the suspected target wildlife; Step S320: Set the visibility score threshold, and calculate the confidence of each key point based on the key point coordinates and visibility score. , when the visibility score is greater than or equal to the visibility score threshold, the confidence of the key point is equal to its visibility score, otherwise the confidence of the key point is equal to 0; The calculation formula of the confidence of the key point is: ,in, is the visibility score threshold; The visibility score threshold is used to filter low-quality key points and is set by those skilled in the art based on experience.

[0033] Step S330: Select the K key points with the highest confidence as the key point features of the suspected target wild animal, forming the feature descriptor of the suspected target wild animal in the t-th frame image. ; The feature descriptor is expressed as: ,in, is the coordinate of the Kth key point, is the confidence of the Kth key point; Figure 2 Flowchart of the target detection method provided for this application; The above steps obtain the key point coordinates and visibility scores of the target wild animals through key point detection, and construct the feature descriptor of the key points based on the confidence. This feature descriptor can be used for subsequent target matching and recognition tasks.

[0034] Step S400: Perform target detection and key point detection on the t+1th frame image to obtain the actual position coordinates and feature descriptors of the candidate target, and perform trajectory matching between the suspected target wild animal and the candidate target. If they match, the actual position coordinates and feature descriptors of the suspected target wild animal are updated; if they do not match, the candidate target is defined as a new target animal; The specific method of performing target detection and key point detection on the t+1th frame image to obtain the actual position coordinates and feature descriptors of the candidate target, performing trajectory matching on the suspected target wild animal and the candidate target, and updating the actual position coordinates and feature descriptors of the suspected target wild animal if they match, and defining the candidate target as a new target animal if they do not match is as follows: Step S410: performing target detection on the t+1th frame image based on the improved YOLOv5 model, outputting the target confidence prediction value, bounding box coordinate prediction value and category probability prediction value of the candidate target in the t+1th frame image, and calculating the actual position coordinates of the candidate target; The method for calculating the actual position coordinates of the candidate target is the same as the method for calculating the actual position coordinates of the suspected target wild animal; Step S420: Perform key point detection on the t+1th frame image to obtain the feature descriptor of the candidate target in the t+1th frame image ; The method for calculating the feature descriptor of the candidate target is the same as the method for calculating the feature descriptor of the suspected target wild animal; Step S430: Calculate the similarity between the candidate target and the suspected target wildlife based on the feature descriptor of the candidate target and the feature descriptor of the suspected target wildlife. ; The calculation formula for the similarity between the candidate target and the suspected target wild animal is: ,in, and are the feature descriptors of the kth key points of the suspected wild animal f and the candidate target g, respectively. represents the L2 norm, represents an exponential function with base e; Step S440: Preset similarity threshold If the similarity between the candidate target and the suspected target wild animal is greater than the similarity threshold, the candidate target and the suspected target wild animal are considered to be matched successfully, and the actual position coordinates of the suspected target wild animal are updated to the actual position coordinates of the candidate target, and the feature descriptor of the suspected target wild animal is updated to the feature descriptor of the candidate target; The similarity threshold is set by those skilled in the art based on experience.

[0035] Step S450: if the similarity between the candidate target and the suspected target wild animal is less than or equal to the similarity threshold, it is considered that the suspected target wild animal is lost in the t+1th frame image, and the candidate target in the t+1th frame image is defined as a new target animal; Figure 3 Flow chart of the key point detection method provided for this application; The above steps achieve cross-frame tracking of wild animals through key point detection and matching, effectively improving the performance of automatic identification and tracking of wild animals, and providing a basis for the subsequent steps to count the total number of target wild animals in the target area.

[0036] Step S500: according to the track matching results of the suspected target wild animal and the candidate target, the identified suspected target wild animal and the new target animal are counted by a counter to obtain the total number of target wild animals in the target area; The specific method of counting the identified suspected target wild animals and new target animals through a counter according to the track matching results of the suspected target wild animals and the candidate targets to obtain the total number of target wild animals in the target area is: Step S510: setting a counter count with an initial value of 0 for counting the total number of target wild animals in the target area; Step S520: traverse all the suspected target wild animals that have been successfully matched, and determine whether the suspected target wild animal has been counted. If the suspected target wild animal has not been counted, add 1 to the value of the counter count; otherwise, the value of the counter count remains unchanged; Step S530: traverse all new target animals to determine whether the new target animal has been counted. If the new target animal has not been counted, add 1 to the value of the counter count; otherwise, the value of the counter count remains unchanged. When the value of the counter count increases by 1, the flag bit of the suspected target wild animal or new target animal is set to true, indicating that it has been counted; Step S540: The final value of the counter count is used as the total number of target wild animals in the target area.

[0037] When the technical solution provided by the present application is actually applied, multiple drones are set up in the target area to simultaneously acquire continuous frame images, and then these continuous frame images are subjected to target detection, key point detection and trajectory matching to ensure that there is no repeated counting, thereby ensuring the accuracy of the target wild animal counting to the greatest extent.

[0038] The above steps are based on the target detection, key point detection and trajectory matching of steps S200 to S400, and by setting a counter, the problem of calculating the number of target wild animals in the total target area is clearly solved.

[0039] Example 2 like Figure 4 As shown, the UAV target intelligent recognition and perception system based on deep learning provided by this application includes: The image monitoring acquisition module is used to divide the target area according to the geographical scope of the wildlife reserve, extract the importance features of each target area and calculate the regional importance score, generate a target area monitoring plan, and the drone acquires continuous frame images based on the target area monitoring plan; The image target detection module is used to perform target detection using the improved YOLOv5 model, input continuous frame images, output the bounding box of the suspected target wild animal in each frame image and map it to the actual geographic coordinate system to obtain the actual position coordinates of the suspected target wild animal in each frame image; The image key point detection module is used to detect key points of the t-th frame image, extract the key point coordinates of the suspected target wild animal, calculate the visibility score of the key point and the confidence of each key point, and the K key points with the highest confidence constitute the feature descriptor; The animal track matching module is used to perform target detection and key point detection on the t+1 frame image to obtain the actual position coordinates and feature descriptors of the candidate target, and to match the track of the suspected target wild animal with the candidate target. If they match, the actual position coordinates and feature descriptors of the suspected target wild animal are updated; if they do not match, the candidate target is defined as a new target animal; The animal number counting module is used to count the suspected target wild animals and new target animals identified through a counter based on the trajectory matching results of the suspected target wild animals and the candidate targets, so as to obtain the total number of target wild animals in the target area.

[0040] Example 3 An embodiment of the present application provides an electronic device for a method for intelligently identifying and perceiving drone targets based on deep learning. The electronic device may include one or more processors and one or more memories. The memory stores a computer-readable code, and when the computer-readable code is executed by one or more processors, the method for intelligently identifying and perceiving drone targets based on deep learning as described above may be executed.

[0041] The electronic device may include a bus, one or more CPUs, a read-only memory (ROM), a random access memory (RAM), a communication port connected to a network, an input / output component, a hard disk, and the like. A storage device in the electronic device, such as a ROM or a hard disk, may store the deep learning-based UAV target intelligent recognition and perception method provided in the present application. The deep learning-based UAV target intelligent recognition and perception method may, for example, include: dividing the target area according to the geographical scope of the wildlife reserve, extracting the importance features of each target area and calculating the regional importance score, generating a target area monitoring plan, and the UAV acquires continuous frame images based on the target area monitoring plan; using an improved YOLOv5 model for target detection, inputting continuous frame images, outputting the bounding box of the suspected target wildlife in each frame image and mapping it to the actual geographic coordinate system, and obtaining the actual position coordinates of the suspected target wildlife in each frame image; performing key point detection on the tth frame image, extracting the key point coordinates of the suspected target wildlife, Calculate the visibility score of the key points and the confidence of each key point, and the K key points with the highest confidence constitute the feature descriptor; perform target detection and key point detection on the t+1 frame image to obtain the actual position coordinates of the candidate target and its feature descriptor, and match the trajectory of the suspected target wild animal with the candidate target. If it matches, update the actual position coordinates and feature descriptor of the suspected target wild animal. If it does not match, define the candidate target as a new target animal; according to the trajectory matching results of the suspected target wild animal and the candidate target, count the identified suspected target wild animals and new target animals through a counter to obtain the total number of target wild animals in the target area. Furthermore, the electronic device may also include a user interface. Of course, the architecture of the above electronic device is only exemplary. When implementing different devices, one or more components in the above electronic device can be omitted according to actual needs.

[0042] Example 4 One embodiment of the present application provides a computer-readable storage medium for a method for intelligent recognition and perception of drone targets based on deep learning. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by a processor, the method for intelligent recognition and perception of drone targets based on deep learning according to the implementation of the present application can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache). Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0043] In addition, according to the implementation mode of the present application, the process described above can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, such as: dividing the target area according to the geographical scope of the wildlife reserve, extracting the importance features of each target area and calculating the regional importance score, generating a target area monitoring plan, and the drone acquires continuous frame images based on the target area monitoring plan; using an improved YOLOv5 model for target detection, inputting continuous frame images, outputting the bounding box of the suspected target wildlife in each frame image and mapping it to the actual geographic coordinate system, and obtaining the actual position coordinates of the suspected target wildlife in each frame image; The t-th frame image is subjected to key point detection, the key point coordinates of the suspected target wild animal are extracted, the key point visibility score and the confidence of each key point are calculated, and the K key points with the highest confidence constitute the feature descriptor; the t+1-th frame image is subjected to target detection and key point detection to obtain the actual position coordinates of the candidate target and its feature descriptor, and the suspected target wild animal is matched with the candidate target by trajectory. If it matches, the actual position coordinates and feature descriptor of the suspected target wild animal are updated; if it does not match, the candidate target is defined as a new target animal; according to the trajectory matching results of the suspected target wild animal and the candidate target, the identified suspected target wild animal and new target animal are counted by a counter to obtain the total number of target wild animals in the target area. When the computer program is executed by the central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0044] The methods, apparatuses, and devices of the present application may be implemented in many ways. For example, the methods, apparatuses, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers recording media storing programs for executing the method according to the present application.

[0045] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0046] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for intelligent recognition and perception of UAV targets based on deep learning, characterized in that: include: Divide the target area according to the geographical scope of the wildlife reserve, extract the importance features of each target area and calculate the regional importance score, generate a target area monitoring plan, and the drone acquires continuous frame images based on the target area monitoring plan; The improved YOLOv5 model is used for target detection. Continuous frame images are input, and the bounding box of the suspected target wild animal in each frame is output and mapped to the actual geographic coordinate system to obtain the actual position coordinates of the suspected target wild animal in each frame. Perform key point detection on the t-th frame image, extract the key point coordinates of the suspected target wild animal, calculate the visibility score of the key point and the confidence of each key point, and the K key points with the highest confidence constitute the feature descriptor; Perform target detection and key point detection on the t+1th frame image to obtain the actual position coordinates and feature descriptors of the candidate target, and match the trajectory of the suspected target wild animal with the candidate target. If they match, update the actual position coordinates and feature descriptors of the suspected target wild animal. If they do not match, define the candidate target as a new target animal. According to the trajectory matching results of the suspected target wild animals and the candidate targets, the identified suspected target wild animals and new target animals are counted by a counter to obtain the total number of target wild animals in the target area.

2. The method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning according to claim 1, characterized in that: The target area is divided according to the geographical scope of the wildlife reserve, the importance features of each target area are extracted and the regional importance score is calculated, and a target area monitoring plan is generated. The specific method for the drone to obtain continuous frame images based on the target area monitoring plan is as follows: Divide the geographical scope of the wildlife reserve into a target areas and assign unique numbers, record the basic information and historical monitoring data of each target area, and establish a target area database. The basic information includes the food abundance, shelter conditions and water availability of the target area, and the area of ​​the target area. The historical monitoring data includes the frequency of occurrence of target wildlife, the population density of target wildlife, the frequency of each behavior pattern of target wildlife being observed, the time of each observation of target wildlife activity, the nearest neighbor distance between the target wildlife and its adjacent wildlife, and the number of target wildlife; Extract the importance features of the target area from the target area database, the importance features include: the frequency of occurrence of target wild animals , Population density of target wildlife and behavioral pattern diversity index , Regional Habitat Quality Index , Time Distribution Regularity Index , spatial distribution law index ; Calculate the importance score of the target area based on the importance features ; The importance scores of the target areas are normalized, and the importance scores of the target areas are mapped to various monitoring parameters according to the preset mapping functions of various monitoring parameters; the monitoring parameters include: monitoring frequency MF, single monitoring duration MD, and number of drones DN; A target area monitoring plan is generated according to the monitoring parameters of each target area, and the UAV obtains continuous frame images according to the target area monitoring plan.

3. The method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning according to claim 2, characterized in that: The improved YOLOv5 model is used for target detection, continuous frame images are input, the bounding box of the suspected target wild animal in each frame image is output and mapped to the actual geographic coordinate system, and the specific method for obtaining the actual position coordinates of the suspected target wild animal in each frame image is: Use the improved YOLOv5 model for wildlife target detection and set the confidence threshold , take continuous frame images as input, perform feature scaling on each input frame image to obtain a feature pyramid, input the feature pyramid into the improved YOLOv5 model, extract the target confidence prediction value, bounding box coordinate prediction value and category probability prediction value corresponding to the prediction bounding box with a confidence greater than the confidence threshold at each size, remove redundant prediction bounding boxes, and convert the target confidence prediction value of the remaining prediction bounding boxes into , bounding box coordinate prediction value and class probability prediction value Output, where Indicates the horizontal offset of the center of the bounding box relative to the upper left corner of the predicted bounding box. Indicates the vertical offset of the center of the bounding box relative to the upper left corner of the predicted bounding box. is the center point of the bounding box, are the bounding box width and the bounding box height, represents the probability that the predicted bounding box belongs to the nth target wildlife category, where n is the total number of target wildlife categories, and the remaining predicted bounding boxes are used as bounding boxes of suspected target wildlife; According to the category probability prediction value of the suspected target wild animal, the target wild animal category corresponding to the largest category probability prediction value is selected as the preliminary identification result; According to the size of the feature map , calculate the target center coordinates on the feature map , map the target center coordinates on the feature map to the continuous frame images, and calculate the normalized coordinates of the target center coordinates in the continuous frame images ; Based on the normalized coordinates and the current GPS coordinates of the drone , UAV image resolution , map the bounding box of the suspected target wild animal to the actual geographic coordinate system, and obtain the actual position coordinates of the suspected target wild animal in each frame of the continuous frame image .

4. The method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning as claimed in claim 3, characterized in that: The specific construction method of the improved YOLOv5 model is: An improved YOLOv5 model is constructed, using CSPDarknet53 as the backbone network. An attention mechanism is added to each stage of the backbone network. The attention mechanism consists of two fully connected layers and a Sigmoid activation function. The input feature map is X and the number of channels is C. The calculation process of the attention mechanism is as follows: perform global average pooling on each channel of X to obtain a feature vector Z with a size of 1×1×C; pass Z through the first fully connected layer FC1 to obtain a feature vector with a size of 1×1× The feature vector U of , where r is the scaling factor, is used to control the parameter amount of the attention mechanism; U is passed through the second fully connected layer FC2 to obtain a feature vector V of size 1×1×C; each element in V is activated by the Sigmoid activation function to obtain the attention weight vector A; the input feature map X is multiplied by the attention weight vector A to obtain the modulated feature map ; The FPN structure is used to fuse feature maps of different sizes, and a feature pyramid is constructed through upsampling and splicing operations; A detection head is set at each dimension of the feature pyramid, wherein the detection head consists of a convolutional layer and a fully connected layer, and the detection head outputs target confidence, bounding box coordinates, and category probability; The confidence loss between the target confidence prediction value and the actual target confidence value is calculated using the binary cross entropy loss function, the bounding box loss between the bounding box coordinate prediction value and the actual bounding box coordinate value is calculated using the GIOU loss function, and the category loss between the category probability prediction value and the actual category probability value is calculated using the multivariate cross entropy loss function. The confidence loss, bounding box loss, and category loss are weighted and summed to obtain the total loss function of the improved YOLOv5 model. Obtain wildlife images from historical times, annotate each wildlife image with the category label and bounding box coordinates of the target wildlife, perform data enhancement on the wildlife images to obtain training images, use the annotated training images as input data, output the corresponding target confidence prediction value, bounding box coordinate prediction value, and category probability prediction value based on the improved YOLOv5 model, and use the value that minimizes the total loss function as the training goal. When the total loss function converges, the training is completed, and a trained improved YOLOv5 model is obtained.

5. The method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning according to claim 4, characterized in that: The specific method of performing key point detection on the t-th frame image, extracting the key point coordinates of the suspected target wild animal, calculating the visibility score of the key point and the confidence of each key point, and forming the feature descriptor with the K key points with the highest confidence is as follows: Based on key point detection technology, the key point coordinates of the suspected target wild animal in the t-th frame image are extracted , the visibility score of the key point is calculated based on the conditional probability of the key point visibility ; Set the visibility score threshold and calculate the confidence of each key point based on the key point coordinates and visibility score , when the visibility score is greater than or equal to the visibility score threshold, the confidence of the key point is equal to its visibility score, otherwise the confidence of the key point is equal to 0; Select the K key points with the highest confidence as the key point features of the suspected target wild animal, and form the feature descriptor of the suspected target wild animal in the t-th frame image. .

6. The method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning according to claim 5, characterized in that: The specific method of performing target detection and key point detection on the t+1th frame image to obtain the actual position coordinates and feature descriptors of the candidate target, performing trajectory matching on the suspected target wild animal and the candidate target, and updating the actual position coordinates and feature descriptors of the suspected target wild animal if they match, and defining the candidate target as a new target animal if they do not match is as follows: Perform target detection on the t+1th frame image based on the improved YOLOv5 model, output the target confidence prediction value, bounding box coordinate prediction value and category probability prediction value of the candidate target in the t+1th frame image, and calculate the actual position coordinates of the candidate target; Perform key point detection on the t+1th frame image to obtain the feature descriptor of the candidate target in the t+1th frame image ; Based on the feature descriptors of the candidate targets and the feature descriptors of the suspected target wild animals, the similarity between the candidate targets and the suspected target wild animals is calculated. ; Preset similarity threshold If the similarity between the candidate target and the suspected target wild animal is greater than the similarity threshold, the candidate target and the suspected target wild animal are considered to be matched successfully, and the actual position coordinates of the suspected target wild animal are updated to the actual position coordinates of the candidate target, and the feature descriptor of the suspected target wild animal is updated to the feature descriptor of the candidate target; If the similarity between the candidate target and the suspected target wild animal is less than or equal to the similarity threshold, the suspected target wild animal is considered to be lost in the t+1th frame image, and the candidate target in the t+1th frame image is defined as a new target animal.

7. The method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning according to claim 6, characterized in that: The specific method of counting the identified suspected target wild animals and new target animals through a counter according to the track matching results of the suspected target wild animals and the candidate targets to obtain the total number of target wild animals in the target area is: Set the counter count, with an initial value of 0, to count the total number of target wild animals in the target area; Traverse all the suspected target wild animals that have been matched successfully, and determine whether the suspected target wild animals have been counted. If the suspected target wild animals have not been counted, add 1 to the value of the counter count; otherwise, the value of the counter count remains unchanged; Traverse all new target animals and determine whether the new target animal has been counted. If the new target animal has not been counted, add 1 to the value of the counter count; Otherwise, the value of the counter count remains unchanged; The final value of the counter count is taken as the total number of target wild animals in the target area.

8. A UAV target intelligent recognition and perception system based on deep learning, which is used to implement the UAV target intelligent recognition and perception method based on deep learning as described in any one of claims 1 to 7, characterized in that: include: The image monitoring acquisition module is used to divide the target area according to the geographical scope of the wildlife reserve, extract the importance features of each target area and calculate the regional importance score, generate a target area monitoring plan, and the drone acquires continuous frame images based on the target area monitoring plan; The image target detection module is used to perform target detection using the improved YOLOv5 model, input continuous frame images, output the bounding box of the suspected target wild animal in each frame image and map it to the actual geographic coordinate system to obtain the actual position coordinates of the suspected target wild animal in each frame image; The image key point detection module is used to detect key points of the t-th frame image, extract the key point coordinates of the suspected target wild animal, calculate the visibility score of the key point and the confidence of each key point, and the K key points with the highest confidence constitute the feature descriptor; The animal track matching module is used to perform target detection and key point detection on the t+1 frame image to obtain the actual position coordinates and feature descriptors of the candidate target, and to match the track of the suspected target wild animal with the candidate target. If they match, the actual position coordinates and feature descriptors of the suspected target wild animal are updated; if they do not match, the candidate target is defined as a new target animal; The animal number counting module is used to count the suspected target wild animals and new target animals identified through a counter based on the trajectory matching results of the suspected target wild animals and the candidate targets, so as to obtain the total number of target wild animals in the target area.

9. An electronic device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the method for intelligent recognition and perception of unmanned aerial vehicle targets based on deep learning as described in any one of claims 1 to 7 are implemented.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which is suitable for being loaded by a processor to execute the steps in the deep learning-based drone target intelligent recognition and perception method as described in any one of claims 1 to 7.

Images