Unmanned aerial vehicle detection method and apparatus

Abstract

The application provides a kind of unmanned aerial vehicle detection method and device, the unmanned aerial vehicle detection method includes: obtaining image to be detected;Based on unmanned aerial vehicle detection model, the image to be detected is detected, and unmanned aerial vehicle detection result is obtained;Wherein, the unmanned aerial vehicle detection model is obtained by training neural network based on the architecture of Transformer with mixed sample set as training sample, with fusion feature as training characteristic;Mixed sample set is obtained by being photographed by multi-dimensional optical equipment in multiple different unmanned aerial vehicle activity scenes;Fusion feature is obtained by cross-modal calculation according to Transformer module by extracting multiple features from mixed sample set respectively.The method of the application can accurately capture the characteristics under each mode, and the fusion feature retains the key features under each mode, and through the complement and enhancement between characteristics, the accuracy of unmanned aerial vehicle detection result in complex environment is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer vision, and in particular to a UAV detection method and device. BACKGROUND

[0002] The extremely small target UAV detection algorithm can improve the detection capability of low-altitude, slow-speed and small flying objects, and help to timely discover and respond to potential air threats. For example, in the field of border security, detecting UAVs can enhance the monitoring of illegal cross-border UAVs and prevent smuggling and smuggling behaviors. In the field of urban security, detecting UAVs helps to timely discover and stop UAVs from illegal filming, delivering contraband, and other behaviors. In addition, in the security work of important activities or places, detecting UAVs can improve the monitoring and defense capability of UAVs and ensure the safety of activities.

[0003] In the prior art, when detecting UAV images using target detection technology, the UAV contour information in a single scene is usually selected as a typical feature for UAV target detection. Due to the diversity of UAVs, the interference of complex backgrounds, and the rapid movement of UAVs, the representation ability of the above detection features is poor, which leads to poor recognition and detection capability of UAVs in complex environments, and further leads to inaccurate UAV detection results. SUMMARY

[0004] The present application provides a UAV detection method and device to solve the problem of poor UAV feature representation in the prior art, which leads to poor recognition and detection capability of UAVs in complex environments, and improves the accuracy of UAV detection results.

[0005] The present application provides a UAV detection method, comprising:

[0006] obtaining an image to be detected;

[0007] detecting the image to be detected based on a UAV detection model to obtain a UAV detection result;

[0008] The UAV detection model is obtained by training a neural network based on the Transformer architecture using a mixed sample set as training samples and fusion features as training features. The mixed sample set is obtained by a multi-dimensional optical device in multiple different UAV activity scenes. The fusion features are obtained by cross-modal calculation of multiple features extracted from the mixed sample set according to the Transformer module.

[0009] According to the UAV detection method provided by the present application, the neural network includes the encoding layer corresponding to the Transformer architecture and the feedforward neural network.

[0010] The unmanned aerial vehicle detection model is specifically obtained through the following steps:

[0011] The sample images in the mixed sample set are normalized to obtain normalized images.

[0012] The unmanned aerial vehicle targets in each channel of the normalized images are labeled to obtain a plurality of channel labeled images.

[0013] The plurality of channel labeled images are feature-encoded based on the encoding layer to obtain encoded features, and the encoded features are interacted and fused according to the feedforward neural network and the Transformer mechanism to obtain the fused features.

[0014] The neural network is iteratively trained according to the fused features, and a loss function value corresponding to the neural network is calculated, and in the case that the loss function value is less than a loss threshold, it is determined that the neural network model training converges, and the unmanned aerial vehicle detection model is obtained.

[0015] According to the unmanned aerial vehicle detection method provided by the application, the mixed sample set is obtained through the following steps:

[0016] The mixed sample set is constructed by the multi-dimensional optical equipment according to the RGB images collected by different filters, a plurality of polarization images of different polarization degrees, near-infrared images, long-wave infrared images and high-dynamic channel images.

[0017] According to the unmanned aerial vehicle detection method provided by the application, the interaction and fusion of the encoded features according to the feedforward neural network and the Transformer mechanism to obtain the fused features comprises:

[0018] The encoded features are flattened, and the flattened features are extracted according to the Transformer architecture based on multiple layers and multiple heads to obtain new encoded features, and the new encoded features and the corresponding flattened features are merged to obtain the fused features.

[0019] According to the unmanned aerial vehicle detection method provided by the application, the neural network adopts a YOLOv9 target detection framework to extract features, and the neural network comprises a multi-feature fusion module and a detection head, and the detection head trains a single size candidate frame for the unmanned aerial vehicle target by adopting bidirectional output.

[0020] According to the unmanned aerial vehicle detection method provided by the application, the normalization of the sample images in the mixed sample set to obtain normalized images comprises:

[0021] The sample images in the mixed sample set are denoised, enhanced, normalized and format-converted to obtain the normalized images.

[0022] The application further provides a UAV detection device, comprising:

[0023] an image acquisition module, configured to acquire an image to be detected;

[0024] an image detection module, configured to detect the image to be detected based on a UAV detection model to obtain a UAV detection result;

[0025] The UAV detection model is obtained by training a neural network based on a Transformer architecture with a mixed sample set as training samples and with fusion features as training features; the mixed sample set is obtained by a multi-dimensional optical device in multiple different UAV activity scenes; and the fusion features are obtained by cross-modal calculation on multiple features extracted from the mixed sample set according to the Transformer module.

[0026] The application further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the UAV detection method according to any one of the above when executing the computer program.

[0027] The application further provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program is executable on a processor to implement the UAV detection method according to any one of the above.

[0028] The application further provides a computer program product comprising a computer program, wherein the computer program is executable on a processor to implement the UAV detection method according to any one of the above.

[0029] The UAV detection method and device provided by the application can accurately capture features in respective modalities, and the fusion features retain key features in respective modalities, and the accuracy of the UAV detection result in a complex environment is improved through complementation and enhancement between features. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0031] Figure 1 is one of flow diagrams of the unmanned aerial vehicle detection method provided by the present application.

[0032] Figure 2 is a flow diagram of the fusion feature acquisition provided by the present application.

[0033] Figure 3 is another flow diagram of the unmanned aerial vehicle detection method provided by the present application.

[0034] Figure 4 is a structural diagram of the unmanned aerial vehicle detection device provided by the present application.

[0035] Figure 5 is a structural diagram of the electronic device provided by the present application. DETAILED DESCRIPTION

[0036] To make the objectives, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0037] The unmanned aerial vehicle detection method and device of the present application will be described below. Figures 1-4

[0038] Figure 1 is one of flow diagrams of the unmanned aerial vehicle detection method provided by the present application, as shown in the figure, the unmanned aerial vehicle detection method comprises the following steps: Figure 1

[0039] Step 110, acquiring an image to be detected.

[0040] In this step, the image to be detected includes a real-time shot image or an image acquired from an unmanned aerial vehicle image database.

[0041] For example, the image to be detected can be a large number of unmanned aerial vehicle images collected in different scenes, and the image content includes diversified environments such as cities, suburbs, mountains and seas.

[0042] ​​In step 120, the image to be detected is detected based on the UAV detection model to obtain a UAV detection result; wherein the UAV detection model is obtained by training a neural network based on a Transformer architecture using a mixed sample set as training samples and using fused features as training features; the mixed sample set is obtained by a multi-dimensional optical device in multiple different UAV activity scenes; and the fused features are obtained by cross-modal calculation of multiple features extracted from the mixed sample set according to a Transformer module.

[0043] In this step, each sample image in the mixed sample set includes a UAV in various poses, distances, and lighting conditions to simulate complex real-world scenes, and the multi-scene UAV image ensures the diversity and richness of the data set.

[0044] In this embodiment, the mixed sample set is obtained by the following steps: constructing the mixed sample set by the multi-dimensional optical device according to the RGB images collected by different filters, the polarization images of multiple different polarization degrees, the near-infrared images, the long-wave infrared images, and the images of high dynamic channels

[0045] In this embodiment, the multi-dimensional optical device can collect UAV images in different scenes and parse them into RGB images, polarization images of three different polarization degrees, long-wave infrared images, near-infrared images, and high-dynamic images of different dimensions.

[0046] In this embodiment, based on different modal images, a feature extraction network can be used to obtain image features in a single modal, a feature fusion module can be used to realize feature interaction and fusion, and a Transformer module can be used to effectively realize cross-modal calculation.

[0047] In this embodiment, by designing a neural network including a network structure with partial weight sharing, the training complexity is reduced, and the Transformer module is allowed to interact information within and between modalities.

[0048] In this embodiment, by using multiple different modal feature extraction networks to extract features of different dimensions from the mixed sample set, the features in each modal can be accurately captured.

[0049] In this embodiment, the trained image features are spliced and sent to the detection head, which inputs into an efficient deep target detection framework to output the final detection result.

[0050] In this embodiment, the neural network uses a YOLOv9 target detection framework to extract features, and the neural network includes a multi-feature fusion module and a detection head. The detection head uses bidirectional output to train a single size candidate box for the UAV target.

[0051] The unmanned aerial vehicle detection method provided by the embodiment of the application can accurately capture features in respective modalities, and the fusion features retain key features in respective modalities, and the accuracy of the unmanned aerial vehicle detection result in a complex environment is improved through complementation and enhancement between features.

[0052] In some embodiments, the neural network comprises an encoding layer corresponding to a Transformer architecture and a feedforward neural network; and the unmanned aerial vehicle detection model is obtained through the following steps: performing normalization processing on each sample image in the mixed sample set to obtain a normalized processing image; marking unmanned aerial vehicle targets in each channel of the normalized processing image to obtain a plurality of channel marked images; performing feature encoding on the plurality of channel marked images based on the encoding layer to obtain encoded features; and performing interaction and fusion on the encoded features according to the feedforward neural network and the Transformer mechanism to obtain fusion features; iteratively training the neural network according to the fusion features, and calculating a loss function value corresponding to the neural network; and in a case where the loss function value is less than a loss threshold, determining that the neural network model converges, and obtaining the unmanned aerial vehicle detection model.

[0053] In this embodiment, a data analysis and marking system applied to multi-dimensional optical data is constructed to extract useful information from multi-dimensional optical data and perform effective data marking; specifically, the data analysis and marking system reads image data of different optical devices, extracts basic information (such as resolution, color space, etc.) of the image, and then marks the target where the unmanned aerial vehicle is located in the multi-dimensional optical image and displays in each channel image.

[0054] In this embodiment, the data analysis and marking system can also support analysis and loading of various optical image formats, and provide a friendly user interface for users to perform image marking.

[0055] For example, the data analysis and marking system can provide rectangular frame, polygon, key point, etc. marking tools, so that the user can accurately mark the unmanned aerial vehicle target in the multi-dimensional optical image, and in the case where part of the channel label is missing, the invention can automatically make up the marking information of other channels.

[0056] In this embodiment, during the image marking process of the data analysis and marking system, the channel images of long-wave infrared are all in an aligned state, because there is a large deviation between the size of the long-wave infrared image and the sizes of other images, and the marking system can assign labels to each channel image through preprocessing.

[0057] In this embodiment, a large number of unmanned aerial vehicle images are collected based on a multi-dimensional optical device, and the collected images are normalized to improve image quality and unify data formats.

[0058] Specifically, the normalized processing of each sample image in the mixed sample set includes denoising, enhancement, normalization and format conversion of each sample image in the mixed sample set to obtain the normalized processing image.

[0059] In this embodiment, after obtaining the normalized processing image, the images are divided into a training set, a validation set and a test set according to certain rules for subsequent network training and evaluation; and finally a multi-dimensional optical unmanned aerial vehicle detection data set is constructed as the basis for network training.

[0060] In this embodiment, the training set, the validation set and the test set can be divided according to a ratio, and the ratio coefficient is set according to user requirements, for example, a plurality of normalized processing images are divided into a training set, a validation set and a test set according to a ratio of 7:2:1.

[0061] In this embodiment, the encoder of the Transformer is used to encode the features of images in different channels, and then the multi-head self-attention mechanism and the feedforward neural network are used to realize effective interaction and fusion between features in different channels.

[0062] In this embodiment, the interaction and fusion of the encoded features according to the feedforward neural network and the Transformer mechanism include: flattening the encoded features, extracting the flattened features according to the multi-layer multi-head Transformer architecture, obtaining new encoded features, and merging the new encoded features and the corresponding flattened features to obtain the fusion features.

[0063] Figure 2 is a flowchart of obtaining fusion features provided by the present application, and in the embodiment shown in Figure 2 In the embodiment shown in the figure, after the image features (input feature 1, input feature 2 and input feature 3) corresponding to the images in different channels are spliced, the features are flattened by the standardization layer, and the features are encoded by combining the multi-head self-attention mechanism to obtain new encoded features, and the new encoded features and the corresponding flattened features are spliced, and finally the spliced features are separated to obtain new output features of different dimensions.

[0064] Specifically, the multi-modal feature fusion based on the Transformer is realized by performing sequence signal feature extraction processing on the image features of each dimension output by the intermediate layer.

[0065] For example, the features of the images of each modality are absorbed and flattened, which are used as sequence information to input a multi-layer multi-head Transformer architecture for feature extraction. The obtained results are split and summed with the original features to input a subsequent feature extraction network for further calculation. Finally, the fusion of the features of multiple different dimensions of modalities is realized, and the corresponding fusion features are obtained.

[0066] In this embodiment, the loss function corresponding to the neural network described above can be a cross-entropy loss function, a positioning loss function, or a combination of the cross-entropy loss function and the positioning loss function.

[0067] Figure 3 Figure 2 is a flowchart of the UAV detection method provided by the present application, in which Figure 3 In the embodiment shown, the data of each sample image in the mixed sample set is analyzed by using a plurality of optical image acquisition devices to obtain five groups of images or data of different modalities, including RGB images, near-infrared images, polarization images, long-wave infrared images, and multi-dimensional optical UAV detection data sets. A pre-trained backbone network (including feature extraction network 1, feature extraction network 2, feature extraction network 3, feature extraction network 4, and feature extraction network 5) is used to extract features from the grouped multi-dimensional images of the five modalities, and then the extracted features are input into a feature fusion network based on a Transformer architecture for cross-channel feature splicing. Finally, a detection head is used to detect the fused features.

[0068] In this embodiment, the trained UAV detection model can be deployed on an algorithm server to perform UAV target detection on the received multi-dimensional optical images by using the multi-modal feature fusion target detection method.

[0069] In this embodiment, the neural network is trained based on a GPU (Graphics Processing Unit, GPU) to obtain a trained UAV detection model (for example, a trained weight file), and the model or file is deployed on an algorithm server to process the received multi-dimensional optical images according to the above process, thereby realizing UAV target detection.

[0070] In this embodiment, the data set inference result shows that the input image is 4096x3000x9, the feature fusion module is used for separate detection of the RGB image, the target detection accuracy after feature fusion is 81.39%, and the target detection accuracy based on the RGB image is only 63.23%, which fully proves the effectiveness of the multi-dimensional optical image feature fusion UAV target detection method.

[0071] The UAV detection method of this embodiment can be freely deployed on an algorithm server and is not limited by the hardware system, thereby maximizing the extraction of data features.

[0072] The unmanned aerial vehicle detection method provided by the embodiment of the application comprises the following steps: performing normalization processing on each sample image in a mixed sample set, marking unmanned aerial vehicle targets in each channel of the normalized processing image to obtain a plurality of channel marked images, performing feature coding on the plurality of channel marked images based on an encoding layer to obtain coded features, and performing interaction and fusion on the coded features based on a feedforward neural network and a Transformer mechanism, and finally performing iterative training on a neural network based on the fused features to obtain an unmanned aerial vehicle detection model.

[0073] The unmanned aerial vehicle detection device provided by the application is described below, and the unmanned aerial vehicle detection device described below can be correspondingly referred to the unmanned aerial vehicle detection method described above.

[0074] Figure 4 FIG. 1 is a structural schematic diagram of the unmanned aerial vehicle detection device provided by the application, as shown in the figure, the unmanned aerial vehicle detection device comprises an image acquisition module 410 and an image detection module 420. Figure 4

[0075] The image acquisition module 410 is configured to acquire an image to be detected.

[0076] The image detection module 420 is configured to detect the image to be detected based on the unmanned aerial vehicle detection model to obtain an unmanned aerial vehicle detection result.

[0077] The unmanned aerial vehicle detection model is obtained by training a neural network based on a Transformer architecture with a mixed sample set as a training sample and with a fused feature as a training feature; the mixed sample set is obtained by photographing a plurality of different unmanned aerial vehicle activity scenes by a multi-dimensional optical device; and the fused feature is obtained by performing cross-modal calculation on a plurality of features extracted from the mixed sample set according to a Transformer module.

[0078] The unmanned aerial vehicle detection device provided by the embodiment of the application can accurately capture features in different modalities, and the fused feature retains key features in each modality, and the accuracy of the unmanned aerial vehicle detection result in a complex environment is improved through the complementation and enhancement between features.

[0079] Figure 5 FIG. 1 is a structural schematic diagram of the electronic device, as shown in the figure, the electronic device comprises a processor 100 and a memory 200. Figure 5 ​As shown, the electronic device can include a processor 510, a communications interface 520, a memory 530, and a communications bus 540, wherein the processor 510, the communications interface 520, and the memory 530 complete mutual communication through the communications bus 540. The processor 510 can invoke a logic instruction in the memory 530 to execute a UAV detection method, which includes: acquiring an image to be detected; detecting the image to be detected based on a UAV detection model to obtain a UAV detection result; wherein the UAV detection model is obtained by training a neural network based on a Transformer architecture with a mixed sample set as training samples and with fusion features as training features; the mixed sample set is obtained by a multi-dimensional optical device in multiple different UAV activity scenes; and the fusion features are obtained by cross-modal calculation of multiple features extracted from the mixed sample set according to a Transformer module.

[0080] In addition, the logic instruction in the memory 530 described above can be implemented in the form of a software function unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0081] On the other hand, the present application also provides a computer program product, which includes a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program is executed by a processor, so that the computer can execute the UAV detection method provided by the above-mentioned methods, which includes: acquiring an image to be detected; detecting the image to be detected based on a UAV detection model to obtain a UAV detection result; wherein the UAV detection model is obtained by training a neural network based on a Transformer architecture with a mixed sample set as training samples and with fusion features as training features; the mixed sample set is obtained by a multi-dimensional optical device in multiple different UAV activity scenes; and the fusion features are obtained by cross-modal calculation of multiple features extracted from the mixed sample set according to a Transformer module.

[0082] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the UAV detection method provided by any of the above methods, and the method comprises: acquiring an image to be detected; and detecting the image to be detected based on a UAV detection model to obtain a UAV detection result; wherein the UAV detection model is obtained by training a neural network based on a Transformer architecture using a mixed sample set as training samples and using fused features as training features; the mixed sample set is obtained by a multi-dimensional optical device in multiple different UAV activity scenes; and the fused features are obtained by performing cross-modal calculation on multiple features extracted from the mixed sample set according to a Transformer module.

[0083] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement it without creative labor.

[0084] From the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be realized by means of software plus a necessary general hardware platform, and of course, it can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0085] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for detecting unmanned aerial vehicles (UAVs), characterized in that, include: Acquire the image to be detected; The image to be detected is analyzed based on the UAV detection model to obtain the UAV detection result; The drone detection model is obtained by training a neural network based on the Transformer architecture using a mixed sample set as training samples and fused features as training features. The mixed sample set is obtained by capturing images of multiple drone activity scenarios using multi-dimensional optical devices. The fused features are obtained by performing cross-modal calculations on multiple features extracted from the mixed sample set using the Transformer module. The neural network includes the encoding layer and the feedforward neural network corresponding to the Transformer architecture; The drone detection model is trained through the following steps: The images of each sample in the mixed sample set are normalized to obtain normalized images; The drone targets in each channel of the normalized image are labeled to obtain multiple channel labeled images; Based on the coding layer, feature encoding is performed on the multiple channel labeled images to obtain coded features; and the coded features are interacted and fused according to the feedforward neural network and Transformer mechanism to obtain the fused features. The neural network is iteratively trained based on the fusion features, and the corresponding loss function value of the neural network is calculated. If the loss function value is less than the loss threshold, the neural network model training is determined to be converged, and the UAV detection model is obtained. The step of interacting and fusing the encoded features based on the feedforward neural network and the Transformer mechanism to obtain the fused features includes: The encoded features are flattened, and the flattened features are extracted according to the multi-layer, multi-head Transformer architecture to obtain new encoded features. The new encoded features and the corresponding flattened features are then merged to obtain the fused features.

2. The UAV detection method according to claim 1, characterized in that, The mixed sample set is obtained through the following steps: The hybrid sample set is constructed using multi-dimensional optical equipment based on RGB images acquired with different filters, multiple polarized images with different degrees of polarization, near-infrared images, long-wave infrared images, and images from high dynamic range channels.

3. The UAV detection method according to claim 1, characterized in that, The neural network uses the YOLOv9 target detection framework to extract features. The neural network includes a multi-feature fusion module and a detection head. The detection head is trained on a single-size candidate box for UAV targets by using bidirectional output.

4. The UAV detection method according to claim 1, characterized in that, The normalization process for each sample image in the mixed sample set to obtain a normalized image includes: The normalized image is obtained by performing denoising, enhancement, normalization, and format conversion on each sample image in the mixed sample set.

5. A drone detection device, employing the drone detection method as described in claim 1, characterized in that, include: The image acquisition module is used to acquire the image to be detected; The image detection module is used to detect the image to be detected based on the UAV detection model to obtain the UAV detection result; The drone detection model is obtained by training a neural network based on the Transformer architecture using a mixed sample set as training samples and fused features as training features. The mixed sample set is obtained by capturing images of multiple drone activity scenarios using multi-dimensional optical devices. The fused features are obtained by performing cross-modal calculations on multiple features extracted from the mixed sample set using the Transformer module.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the UAV detection method as described in any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the UAV detection method as described in any one of claims 1 to 4.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the UAV detection method as described in any one of claims 1 to 4.

Images