An unmanned aerial vehicle state detection image recognition method, device, equipment and medium

Abstract

This invention belongs to the field of image detection, specifically disclosing a method, device, equipment, and medium for UAV state detection image recognition. The invention acquires images via a UAV and transmits them to the cloud; the cloud receives the images and preprocesses them; the preprocessed images undergo feature recognition and processing to obtain image recognition results; by deeply analyzing the characteristics of line discharge images and utilizing image processing technology, it simply and ingeniously achieves automatic recognition of line discharge images; compared with deep learning algorithms, it does not require complex deep learning models to achieve the target function, saving significant manpower and material costs; the recognition accuracy is high, reaching over 98%, approaching that of deep learning recognition methods.

Description

Technical Field

[0001] This invention belongs to the field of image detection, specifically relating to a method, apparatus, device, and medium for image recognition of UAV status detection. Background Technology

[0002] With the increasing prevalence of drone technology in the power industry, power line inspection is rapidly developing towards intelligence, efficiency, and convenience. Drone inspection technology offers several key advantages: First, it provides strong real-time performance. Drones can be equipped with high-definition cameras and sensors to detect line defects and anomalies in real time and transmit data, facilitating timely problem identification and resolution. Second, it offers high safety. Drone inspections can be conducted at high altitudes or in complex environments, avoiding the safety risks associated with manual inspections. Third, it is cost-effective. The cost of drone inspections mainly includes equipment purchase and maintenance, and operator salaries, making it significantly lower than traditional manual inspections. Fourth, it is not limited by terrain. Drones can easily traverse complex terrain and dangerous areas for inspection, where manual inspections may pose significant safety hazards. Fifth, it is reusable. Drones can be recovered and reused after missions, improving equipment utilization. Sixth, it enables intelligent management. Drone inspections can be integrated with intelligent management platforms to achieve automatic data processing, analysis, and early warning functions, improving management efficiency and accuracy.

[0003] Currently, when using drones for autonomous power line inspection and image recognition, the drones are equipped with high-definition cameras and sensors, and fly along power lines via remote control or autonomous flight mode. Depending on the flight settings, the captured information is uploaded in real time to a ground station or data center for analysis and processing (or temporarily stored in the drone's internal memory and copied after the flight). Ground station or data center personnel analyze the data transmitted by the drones to determine if there are any faults or safety hazards in the lines, and formulate corresponding maintenance and repair plans.

[0004] Currently, defect identification in UAV condition detection images mainly relies on manual methods or models built using deep learning. Manual identification suffers from inconsistent accuracy and high labor costs. While model algorithms are highly efficient, they require databases and hardware such as GPUs (top-tier GPUs are subject to export restrictions to my country), thus demanding significant human and material resources. Summary of the Invention

[0005] The purpose of this invention is to provide a method, apparatus, device, and medium for UAV status detection image recognition, in order to solve the problems of unstable accuracy and high labor costs in existing UAV status monitoring image recognition methods.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for image recognition of unmanned aerial vehicle (UAV) state detection, comprising:

[0008] The drone acquires images and transmits them to the cloud;

[0009] Images are received and preprocessed in the cloud.

[0010] The preprocessed image is subjected to feature recognition and processing to obtain the image recognition result.

[0011] Furthermore, the transmission of images acquired by the drone to the cloud specifically involves:

[0012] The drone is set to travel along a pre-set route. The drone is equipped with a power status detection device and performs inspections along the pre-set route. It then takes pictures at the designated location. After taking pictures, the drone transmits the images to the cloud in real time via wireless signal, or stores the pictures on the drone's built-in storage card. After the drone completes its flight mission, the images are copied to the cloud for further processing.

[0013] Furthermore, after receiving the image captured by the drone, the cloud performs noise reduction processing using a Gaussian filtering method. After noise reduction, the image is then masked to obtain a pre-processed image.

[0014] Furthermore, the masking process specifically includes:

[0015] The image is converted from RGB format to HSV format, and then processed using masks of different colors.

[0016] In HSV format, when using blue for masking, the blue mask parameters are: H ranges from [100, 124], S ranges from [43, 255], and V ranges from [43, 255].

[0017] When yellow is used for masking, the yellow mask parameters are: H ranges from [26, 77], S ranges from [43, 255], and V ranges from [43, 255].

[0018] When using other colors for masking, set them according to the HSV value range.

[0019] Furthermore, the step of performing feature recognition and processing on the preprocessed image to obtain the image recognition result specifically includes:

[0020] When using blue and yellow masks to perform feature recognition and processing on the preprocessed image:

[0021] The Hough circular transform method is used for feature recognition of images;

[0022] Feature processing is performed on the image after feature recognition to obtain the image recognition result.

[0023] Furthermore, the method of using Hough circular transform to perform feature recognition on the image specifically includes:

[0024] When using blue and yellow masks to perform feature recognition and processing on the preprocessed image:

[0025] The Hough circular transform method is used to detect the preprocessed image and obtain the center coordinates and radius information of the feature circle;

[0026] Set the parameters for the Hough circular transform:

[0027] dp=1, minDist=1000, param1=10, param2=10, minRadius=1, maxRadius=100;

[0028] In the formula, dp is the accumulator resolution; minDist is the minimum distance between the center points; param1 is the high threshold of the Canny edge detector; param2 is the number of votes that the center point must receive; minRadius is the minimum radius of the detection circle; and maxRadius is the maximum radius of the detection circle.

[0029] Image feature values ​​under the blue mask, Rb={(xb i yb i rb i The data size is 3n, where i = 1, 2, ..., n.

[0030] Image feature values ​​under the yellow mask, Ry={(xy j yy j ry j )|j=1,2,…m},data volume is 3m.

[0031] Furthermore, the feature processing of the image after feature recognition to obtain the image recognition result specifically includes:

[0032] Determine the relative positions of the feature circles under the blue and yellow masks;

[0033] When the yellow ring is contained within the blue ring, the image shows a discharge phenomenon; otherwise, the image does not contain discharge information.

[0034] The yellow circle is completely contained within the blue circle, based on the following data:

[0035] d ij <rbi -ry j

[0036] In the formula, d ij rb is the circular distance between the i-th circle under the blue mask and the j-th circle under the yellow mask; i Let y be the radius of the i-th circle under the blue mask; j Let be the radius of the j-th circle in the yellow mask;

[0037] When the above judgment conditions are met, it is determined that there is a discharge phenomenon in the image, and the image recognition result is obtained.

[0038] In a second aspect, the present invention provides a drone state detection image recognition device, comprising:

[0039] The image acquisition module is used to acquire images via drone and transmit them to the cloud;

[0040] The preprocessing module is used to receive images from the cloud and preprocess them.

[0041] The recognition module is used to perform feature recognition and processing on the preprocessed image to obtain the image recognition result.

[0042] In a third aspect, the present invention provides an electronic device including a processor and a memory, the processor being configured to execute a computer program stored in the memory to implement a UAV state detection image recognition method as described in any of the preceding claims.

[0043] In a fourth aspect, the present invention provides a computer-readable storage medium storing at least one instruction that, when executed by a processor, implements a UAV state detection image recognition method as described in any one of the preceding claims.

[0044] The beneficial effects of this invention are as follows:

[0045] 1. This invention acquires images via drones and transmits them to the cloud; the cloud receives the images and preprocesses them; the preprocessed images undergo feature recognition and processing to obtain image recognition results; by deeply analyzing the characteristics of line discharge images and utilizing image processing technology, automatic recognition of line discharge images is achieved simply and ingeniously; compared with deep learning algorithms, it does not require complex deep learning models to achieve the target function, saving a significant amount of manpower and material resources; the recognition accuracy is high, reaching over 98%, approaching that of deep learning recognition methods.

[0046] 2. This invention has low hardware requirements. It does not need to run on a high-performance processor; a regular processing unit can meet the functional requirements, which can save a lot of hardware costs. Starting from the features of the image itself, the target function can be achieved without building an image library. It is highly portable. Because the implementation process is relatively simple and the program is small, it can be widely used in various hardware devices, such as being directly mounted on a drone to achieve real-time recognition. Attached Figure Description

[0047] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0048] Figure 1 This is a schematic diagram of a method for image recognition based on the state detection of a drone.

[0049] Figure 2 These are images taken by a drone after it was equipped with a condition monitoring device.

[0050] Figure 3 This is a photo of an image after filtering.

[0051] Figure 4 Image after processing with a blue mask;

[0052] Figure 5 The first result is the identification of blue mask features without optimization.

[0053] Figure 6 The result is the identification of the second type of blue mask feature without optimization.

[0054] Figure 7 The optimized blue mask feature recognition results;

[0055] Figure 8 This is a structural block diagram of a drone state detection and image recognition device.

[0056] Figure 9 This is a structural block diagram of an electronic device. Detailed Implementation

[0057] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0058] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0059] Example 1

[0060] like Figure 1 As shown, a method for UAV state detection image recognition includes:

[0061] S1: The drone acquires images and transmits them to the cloud;

[0062] Set the drone's walking route. The drone, equipped with a power status detection device, will patrol along the pre-set route and take pictures when it reaches the designated location. After taking pictures, the drone can transmit the images to the cloud in real time via wireless signal, or store the captured images in the drone's built-in storage card. After the drone completes its flight mission, the images will be copied to the cloud for further processing.

[0063] During the shooting process, the drone will automatically adjust the focus according to the shooting situation to ensure that the photos are clear and undistorted. Figure 2 As shown;

[0064] S2: The cloud receives images and preprocesses them;

[0065] Specifically, after receiving images captured by drones in the cloud, the images are first denoised using a Gaussian filter, and then masked to obtain pre-processed images.

[0066] S21: Images captured by drones typically contain some random noise. Therefore, Gaussian filtering is used to denoise the images, resulting in a denoised image, such as... Figure 3 As shown;

[0067] S22: Perform masking (image thresholding) on ​​the denoised image to obtain the preprocessed image;

[0068] To identify cloud patterns in images captured by drones (cloud patterns will appear in the image if there is a discharge in the circuit), Figure 3 The image (enclosed by a Chinese box) is processed using a mask method. The resulting image is shown below. Figure 4 As shown; typically, cloud maps contain more than two colors, but this invention uses two masks, blue and yellow;

[0069] After obtaining the preprocessed image, it can be found that the image with the discharge phenomenon will contain one or more circles; the subsequent judgment method of this invention is mainly based on this "feature circle"; in fact, in order to obtain the discharge intensity of the discharge point, the charged detection instrument usually focuses, and at this time the cloud map will show a circular distribution.

[0070] from Figure 2 Cloud image analysis shows that the blue circle is on the outermost layer, and the yellow circle is inside the blue circle. Therefore, this feature will be used to analyze the discharge information in the image.

[0071] from Figure 3 As can be seen, if there is a discharge in the circuit, the image captured by the drone will contain a cloud pattern, which typically contains at least two colors: blue and yellow. The number of colors is related to the discharge intensity; as the discharge intensity increases, the number of colors contained in the cloud pattern will gradually increase, such as... Figure 3 The image also contains green and red. To perform masking, we need to convert the image from RGB to HSV format, and then use masks of different colors to process the image. Looking up the table, in HSV format, the parameters for the blue mask are: H values ​​range [100, 124], S values ​​range [43, 255], and V values ​​range [43, 255]; the parameters for the yellow mask are: H values ​​range [26, 77], S values ​​range [43, 255], and V values ​​range [43, 255]. If other color masks are needed, they can be set according to the HSV value range. The image after blue masking is shown below. Figure 4 As shown.

[0072] S3: Perform feature recognition and processing on the preprocessed image to obtain the image recognition result;

[0073] To obtain the feature information (i.e., feature circle information) of the blue and yellow masks, the Hough circular transform method is used for image recognition. This method can obtain the information of the feature circles, namely the center coordinates and radius. Then, by determining the relative positions of the feature circles under the blue and yellow masks, if there is a case where the yellow circle is contained within the blue circle, we consider that the image contains a discharge phenomenon; otherwise, the image does not contain discharge information.

[0074] Specifically,

[0075] S31: Perform feature recognition on the preprocessed image to obtain all features of the blue and yellow masks;

[0076] like Figure 4As shown, a ring (within the blue box) exists in the image after blue mask processing. This feature can serve as an important basis for identifying whether the image contains discharge information (in fact, in charged detection, the image needs to be focused to determine the discharge intensity, at which point the cloud map will become circular). Therefore, in order to extract... Figure 3 The Hough circular rings in the feature rings are detected using the Hough circular ring transform method in this invention.

[0077] Using this method, we can obtain information about all circles in the image after blue mask processing, namely the center coordinates and radius of each circle. This information about the circles under the blue mask is called the image feature value under the blue mask, Rb={(xb i yb i rb i The data size is 3n, where i = 1, 2, ..., n.

[0078] Similarly, using the yellow mask, we can obtain the image feature values ​​under the yellow mask, Ry={(xy j yy j ry j The data size is 3m, where j = 1, 2, ..., m.

[0079] In fact, when using the Hough circular transform method for image detection, if the parameters are not selected properly, it may detect a large amount of noisy circular information or fail to detect anything at all. Figure 5 and Figure 6 The screenshot shows the detection results before optimization. To reduce computation and improve detection accuracy, the parameters of the Hough circular transform were optimized. One reasonable value is:

[0080] dp=1, minDist=1000, param1=10, param2=10, minRadius=1, maxRadius=100;

[0081] In the formula, dp is the accumulator resolution; minDist is the minimum distance between the center points; param1 is the high threshold of the Canny edge detector; param2 is the number of votes that the center point must receive; minRadius is the minimum radius of the detection circle; and maxRadius is the maximum radius of the detection circle.

[0082] Figure 7 The results of image detection using the above parameters are shown. It can be seen that the desired target circle has been detected, and its feature values ​​under the blue mask are obtained: (358.5 270.5 45.2); similarly, the feature values ​​under the yellow mask can be obtained using the same parameters: (358.5 269.5 35.5).

[0083] Obtain all features of the blue and yellow masks;

[0084] S32: Perform feature processing on the image after feature recognition to obtain the image recognition result;

[0085] After obtaining all features of the blue and yellow masks, the following analysis is performed: if there is a yellow ring completely contained within a blue ring, there is a discharge cloud map in the image, that is, there is a discharge phenomenon in the detected line segment.

[0086] The specific determination process is as follows:

[0087] To determine whether the yellow circle is completely contained within the blue circle, the following data is used:

[0088] d ij <rb i -ry j

[0089] In the formula, d ij rb is the circular distance between the i-th circle under the blue mask and the j-th circle under the yellow mask; i Let y be the radius of the i-th circle under the blue mask; j Let be the radius of the j-th circle in the yellow mask;

[0090] In fact, after we have optimized the Hough circle detection parameters, the number of circles that the two masks can detect is very limited. Therefore, the number of judgments for the above formula is less, which also illustrates the necessity of optimizing the Hough circle detection parameters.

[0091] When a value matching the above formula appears, we consider that a cloud map exists in the image, indicating a discharge phenomenon. Based on the blue and yellow mask data calculated above, we calculate:

[0092]

[0093] When the above judgment conditions are met, it is determined that there is a discharge phenomenon in the image, and the image recognition result is obtained.

[0094] Example 2

[0095] like Figure 8 As shown, based on the same inventive concept as the above embodiments, the present invention also provides a UAV state detection image recognition device, comprising:

[0096] The image acquisition module is used to acquire images via drone and transmit them to the cloud;

[0097] The preprocessing module is used to receive images from the cloud and preprocess them.

[0098] The recognition module is used to perform feature recognition and processing on the preprocessed image to obtain the image recognition result.

[0099] Example 3

[0100] like Figure 9 As shown, the present invention also provides an electronic device 100 for implementing a method for detecting and recognizing images of a drone's state;

[0101] The electronic device 100 includes a memory 101, at least one processor 102, a computer program 103 stored in the memory 101 and executable on at least one processor 102, and at least one communication bus 104.

[0102] The memory 101 can be used to store computer program 103. The processor 102 implements the steps of the UAV state detection image recognition method of Embodiment 1 by running or executing the computer program stored in the memory 101 and calling the data stored in the memory 101.

[0103] The memory 101 may primarily include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the electronic device 100 (such as audio data), etc. In addition, the memory 101 may include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0104] At least one processor 102 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 102 may be a microprocessor or any conventional processor. Processor 102 is the control center of electronic device 100, connecting various parts of electronic device 100 via various interfaces and lines.

[0105] The memory 101 in the electronic device 100 stores multiple instructions to implement a UAV state detection image recognition method, and the processor 102 can execute multiple instructions to achieve the following:

[0106] The drone acquires images and transmits them to the cloud;

[0107] Images are received and preprocessed in the cloud.

[0108] The preprocessed image is then subjected to feature recognition and processing to obtain the image recognition result;

[0109] Example 4

[0110] If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, and read-only memory (ROM).

[0111] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0112] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0113] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0114] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0115] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not 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 modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for image recognition of unmanned aerial vehicle (UAV) status detection, characterized in that, include: The drone acquires images and transmits them to the cloud; Images are received and preprocessed in the cloud. Preprocessing is done using a mask. Masking processing includes: converting the image from RGB format to HSV format, and processing the image using masks of different colors; The preprocessed image is subjected to feature recognition and processing to obtain image recognition results, including: when the preprocessed image is subjected to feature recognition and processing using blue and yellow masks, the Hough circular transform method is used to perform feature recognition on the image; Feature processing is performed on the image after feature recognition to obtain the image recognition result, including: determining the relative positions of the feature circles under the blue and yellow masks; when the yellow ring is contained within the blue ring, the image exhibits a discharge phenomenon; otherwise, the image does not contain discharge information; the yellow circle is completely contained within the blue circle, based on the following data: In the formula, Let be the circular distance between the i-th circle under the blue mask and the j-th circle under the yellow mask; Let be the radius of the i-th circle under the blue mask; Let be the radius of the j-th circle of the yellow mask; when the above judgment condition is met, it is determined that there is a discharge phenomenon in the image, and the image recognition result is obtained.

2. The UAV state detection image recognition method according to claim 1, characterized in that, The specific steps involved in transmitting the images acquired by the drone to the cloud are as follows: The drone is set to travel along a pre-set route. The drone is equipped with a power status detection device and performs inspections along the pre-set route. It then takes pictures at the designated location. After taking pictures, the drone transmits the images to the cloud in real time via wireless signal, or stores the pictures on the drone's built-in storage card. After the drone completes its flight mission, the images are copied to the cloud for further processing.

3. The UAV state detection image recognition method according to claim 1, characterized in that, In the process of receiving and preprocessing images in the cloud, before masking the images, the images are first denoised using a Gaussian filtering method.

4. The UAV state detection image recognition method according to claim 3, characterized in that, Image processing is performed using masks of different colors, including: In HSV format, when blue is used for masking, the blue mask parameters are: H ranges from [100, 124], S ranges from [43, 255], and V ranges from [43, 255]. When yellow is used for masking, the yellow mask parameters are: H ranges from [26, 77], S ranges from [43, 255], and V ranges from [43, 255]. When using other colors for masking, set them according to the HSV value range.

5. The UAV state detection image recognition method according to claim 1, characterized in that, The method of using Hough circular transform for image feature recognition specifically includes: When performing feature recognition and processing on preprocessed images using blue and yellow masks: The Hough circular transform method is used to detect the preprocessed image and obtain the center coordinates and radius information of the feature circle; Set the parameters for the Hough circular transform: dp=1, minDist=1000, param1=10, param2=10, minRadius=1, maxRadius=100; In the formula, dp is the accumulator resolution; minDist is the minimum distance between the center points; param1 is the high threshold of the Canny edge detector; param2 is the number of votes that the center point must receive; minRadius is the minimum radius of the detection circle; and maxRadius is the maximum radius of the detection circle. Image feature values ​​under the blue mask The data volume is 3n; Image feature values ​​under the yellow mask The data size is 3MB.

6. A UAV state detection image recognition device, used to implement the UAV state detection image recognition method as described in any one of claims 1 to 5, characterized in that, include: The image acquisition module is used to acquire images via drone and transmit them to the cloud; The preprocessing module is used to receive images from the cloud and preprocess them. The recognition module is used to perform feature recognition and processing on the preprocessed image to obtain the image recognition result.

7. An electronic device, characterized in that, It includes a processor and a memory, the processor being used to execute a computer program stored in the memory to implement a UAV state detection image recognition method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one instruction, which, when executed by a processor, implements a UAV state detection image recognition method as described in any one of claims 1 to 5.

Images