Image rapid identification method and system based on unmanned aerial vehicle

Abstract

The invention discloses an image rapid identification method and system based on an unmanned aerial vehicle, the unmanned aerial vehicle comprises a first shooting assembly and a second shooting assembly, the method comprises the following steps: the first shooting assembly of the unmanned aerial vehicle obtains a first image of a road, and the first image comprises a target object; identifying position information of a target object in the first image, and determining a shooting position and focal length information of the unmanned aerial vehicle according to the position information; the unmanned aerial vehicle controls the second shooting assembly to adjust the focal length according to the focal length information at the shooting position and then obtains a second image of each target object in the first image; determining a first area of the target object in the first image, replacing an image of the first area according to the second image, determining a to-be-uploaded target image, uploading the target image to a server, receiving the target image by the server, analyzing the target object by adopting an analysis model based on a tensor processor, and determining an analysis result.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and system for rapid image recognition based on unmanned aerial vehicles. Background Art

[0002] Existing road monitoring usually involves setting up cameras on the road, collecting images of vehicles on the road by taking photos, and uploading them to a server for analysis.

[0003] However, the position of the camera in this solution is fixed, and the images of some vehicles may not be clear. Summary of the invention

[0004] The present invention provides a method and system for rapid image recognition based on an unmanned aerial vehicle, which can obtain clearer vehicle images for analysis.

[0005] In order to solve the above-mentioned technical problems, the present invention is achieved as follows: In a first aspect, the present application provides a method for rapid image recognition based on a drone, wherein the drone comprises a first shooting component and a second shooting component, wherein the first shooting component is used to obtain a first image of a road, and the second shooting component is used to focus on target objects in the first image respectively to obtain a second image of the target object, wherein the method comprises: the first shooting component of the drone obtains a first image of the road, wherein the first image contains the target object; identifying position information of the target object in the first image, and determining the shooting position and focal length information of the drone based on the position information; the drone controls the second shooting component at the shooting position to adjust the focal length according to the focal length information, and obtains a second image of each target object in the first image; determining a first area of ​​the target object in the first image, and replacing the image of the first area based on the second image, determining the target image to be uploaded, and uploading the target image to a server, wherein the server receives the target image, and adopts an analysis model based on a tensor processor to analyze the target object and determine the analysis result.

[0006] Preferably, the method also includes: when the server determines that a violation has occurred on the road based on the analysis results, it generates time period information and event location, and sends it to the drone; the drone receives the time period information, obtains the first video data shot by the first shooting component, and intercepts the target video data based on the time period information, and uploads it to the server; the drone adjusts its position according to the event location, and shoots the second video data based on the first shooting component, and uploads it to the server.

[0007] Preferably, the drone controls the second shooting component at the shooting position to adjust the focal length according to the focal length information, and then obtains a second image of each target object in the first image, including: the drone controls the second shooting component at the shooting position to adjust the focal length according to the focal length information, and then obtains a first intermediate image; the drone controls the second shooting component at the shooting position to focus on the target object, and then obtains a second intermediate image; and compares the clarity of the first intermediate image and the second intermediate image to obtain the second image.

[0008] Preferably, the drone controls the second shooting component at the shooting position to focus on the target object and obtain a second intermediate image, including: the drone moves to the shooting position and determines the focal length range according to the focal length information; obtains multiple acquired images within the focal length range at the acquisition interval, and compares the clarity of the multiple acquired images, so as to reduce the focal length range and the acquisition interval according to the focal length of the clearest acquired image, so as to obtain the clearest acquired image as the second intermediate image, and the step of comparing the image clarity is determined by comparing the brightness values ​​of a first pixel of the target object and a second pixel adjacent to the first pixel in the acquired image.

[0009] Preferably, the areas photographed by the first shooting component and the second shooting component of the drone at least partially overlap, and the method also includes: the drone obtains a first environmental video related to the road through the first shooting component at the environment collection position, and obtains a second environmental video related to the road through the second shooting component, the first shooting component shoots the video with a first shooting focal length, and the second shooting component shoots the video with a second shooting focal length; the drone uploads the first video data, the first shooting focal length, the second video data, the second shooting focal length and the environment collection position of the drone to the server for modeling on the server to form a road model.

[0010] Preferably, a first area of ​​a target object in a first image is determined, and an image of the first area is replaced based on a second image, and a target image to be uploaded is determined so as to upload the target image to a server, including: determining a first area containing the target object and a second area not containing the target object in the first image; cropping an object image containing the target object from the second image, and replacing the image of the first area based on the object image, and uploading the image to the server; the drone uploads parameter information when the first shooting component shoots the first image to the server, and extracts pixel values ​​of a plurality of fixed pixels from the second area of ​​the first image, and uploads the information to the server, the server extracts a background image from the environment model based on the parameter information, renders the background image based on the pixel values ​​of the plurality of fixed pixels, and performs image restoration in combination with the object image to obtain a restored target image, wherein the parameter information includes the position of the drone and the orientation of the first shooting component.

[0011] Preferably, the server receives the target image and adopts the tensor processor-based analysis model to analyze the target object, and the steps of determining the analysis results include: the server receives the target image and obtains road-related noise-related parameters, and the noise-related parameters include weather information; when it is determined that the road contains noise based on the noise-related parameters, the server obtains the relevant image of the target image from the drone, and denoises the target image based on the target image and the relevant image to obtain a denoised image, and the relevant image of the target image includes images before or after the target image; the denoised image is input into the tensor processor-based analysis model for rapid analysis to determine the analysis result, and the analysis result includes a violation analysis result or a vehicle information analysis result.

[0012] In the second aspect, the present application provides an image rapid recognition system based on a drone, wherein the drone includes a first shooting component and a second shooting component, the first shooting component is used to acquire a first image of a road, and the second shooting component is used to focus on target objects in the first image respectively to acquire a second image of the target object, the system includes: a first image acquisition module, used to acquire a first image of the road through the first shooting component of the drone, the first image including the target object; a position information recognition module, used to identify the position information of the target object in the first image, and determine the shooting position and focal length information of the drone based on the position information; a second image acquisition module, used to control the second shooting component at the shooting position of the drone to adjust the focal length according to the focal length information, and acquire the second image of each target object in the first image; a target image determination module, used to determine a first area of ​​the target object in the first image, and replace the image of the first area based on the second image, determine the target image to be uploaded, and upload the target image to the server, the server receives the target image, and adopts an analysis model based on a tensor processor to analyze the target object and determine the analysis result.

[0013] In a third aspect, the present application provides an electronic device comprising: a memory and at least one processor; the memory is used to store computer-executable instructions; the at least one processor is used to execute the computer-executable instructions stored in the memory, so that the at least one processor executes the method described in the first aspect.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0015] This application can be applied in the scenario of acquiring road images and analyzing the images. This solution can use the movement of the drone to acquire a clearer image of the target object (vehicle) on the road, and upload the clear image to the server for analysis. The drone may include a first shooting component and a second shooting component. The first shooting component is used to acquire a first image of the road, and the second shooting component is used to focus on the target object in the first image respectively to acquire a second image of the target object. The drone can acquire the first image through the first shooting component, and use the long-distance focusing ability of the second shooting component to acquire a clear second image of the target object, and then merge the first image and the second image and upload them to the server. This solution can obtain a clearer vehicle image for analysis. Specifically, the first image of the road can be obtained through the first shooting component of the drone, and the first image contains the target object; then the position information of the target object in the first image is identified, and the shooting position and focal length information of the drone are determined based on the position information; the drone controls the second shooting component at the shooting position to adjust the focal length according to the focal length information, and obtains the second image of each target object in the first image; the first area of ​​the target object in the first image is determined, and the image of the first area is replaced based on the second image, and the target image to be uploaded is determined, and the target image is uploaded to the server, and the server receives the target image to analyze the target object using an analysis model based on a tensor processor to determine the analysis result. This solution can use a tensor processor (TPU) to quickly implement image processing and analysis to determine violation analysis results and vehicle information analysis results, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 This is a flowchart of a method for rapid image recognition based on a drone according to an embodiment of the present application; Figure 2 It is a structural diagram of a drone-based rapid image recognition system according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] This application can be applied in the scenario of acquiring road images and analyzing the images. This solution can use the movement of the drone to acquire a clearer image of the target object (vehicle) on the road, and upload the clear image to the server for analysis. The drone may include a first shooting component and a second shooting component. The first shooting component is used to acquire a first image of the road, and the second shooting component is used to focus on the target object in the first image respectively to acquire a second image of the target object. The drone can acquire the first image through the first shooting component, and use the long-distance focusing ability of the second shooting component to acquire a clear second image of the target object, and then merge the first image and the second image and upload them to the server. This solution can obtain a clearer vehicle image for analysis. Specifically, the first image of the road can be obtained by the first shooting component of the drone, and the first image contains the target object; then the position information of the target object in the first image is identified, and the shooting position and focal length information of the drone are determined according to the position information; the drone controls the second shooting component to adjust the focal length according to the focal length information at the shooting position, and obtains the second image of each target object in the first image; the first area of ​​the target object in the first image is determined, and the image of the first area is replaced according to the second image, and the target image to be uploaded is determined, so as to upload the target image to the server, and the server receives the target image, and adopts the analysis model based on the tensor processor to analyze the target object and determine the analysis result. The shooting position in this scheme can be the position where the drone shoots the first image, or it can be the position after moving based on the position where the first image is shot. This scheme focuses on the position of the target object, and will not focus on objects that are too close or too far away in the case of noise interference (such as rain, snow, fog, etc.), so that a clearer image can be obtained. This scheme can use the tensor processor to quickly realize image processing and analysis to determine the violation analysis results and vehicle information analysis results.

[0019] Specifically, an embodiment of the present application provides a method for rapid image recognition based on a drone, wherein the drone includes a first shooting component and a second shooting component, wherein the first shooting component is used to obtain a first image of a road, and the second shooting component is used to focus on a target object in the first image to obtain a second image of the target object, such as Figure 1 As shown, the method includes: Step 102: A first shooting component of the drone acquires a first image of the road, where the first image contains a target object.

[0020] Step 104: Identify the position information of the target object in the first image, and determine the shooting position and focal length information of the drone based on the position information.

[0021] Step 106: The drone controls the second shooting component at the shooting position to adjust the focal length according to the focal length information, and then obtains a second image of each target object in the first image.

[0022] Step 108, determine the first area of ​​the target object in the first image, and replace the image of the first area according to the second image, determine the target image to be uploaded, upload the target image to the server, the server receives the target image, and adopts the tensor processor-based analysis model to analyze the target object and determine the analysis result.

[0023] This application can be applied in the scenario of acquiring road images and analyzing the images. This solution can use the movement of the drone to acquire a clearer image of the target object (vehicle) on the road, and upload the clear image to the server for analysis. The drone may include a first shooting component and a second shooting component. The first shooting component is used to acquire a first image of the road, and the second shooting component is used to focus on the target object in the first image respectively to acquire a second image of the target object. The drone can acquire the first image through the first shooting component, and use the long-distance focusing ability of the second shooting component to acquire a clear second image of the target object, and then merge the first image and the second image and upload them to the server. This solution can obtain a clearer vehicle image for analysis. Specifically, the first image of the road can be obtained by the first shooting component of the drone, and the first image contains the target object; then the position information of the target object in the first image is identified, and the shooting position and focal length information of the drone are determined according to the position information; the drone controls the second shooting component to adjust the focal length according to the focal length information at the shooting position, and obtains the second image of each target object in the first image; the first area of ​​the target object in the first image is determined, and the image of the first area is replaced according to the second image, and the target image to be uploaded is determined, so as to upload the target image to the server, and the server receives the target image, and adopts the analysis model based on the tensor processor to analyze the target object and determine the analysis result. The shooting position in this scheme can be the position where the drone shoots the first image, or it can be the position after moving based on the position where the first image is shot. This scheme focuses on the position of the target object, and will not focus on objects that are too close or too far away in the case of noise interference (such as rain, snow, fog, etc.), so that a clearer image can be obtained. This scheme can use the tensor processor to quickly realize image processing and analysis to determine the violation analysis results and vehicle information analysis results.

[0024] When a violation occurs, the server can send instructions to the drone, and the drone can upload historical videos related to the violation time according to the instructions, and adjust the shooting position to obtain videos of the area related to the violation event and upload them. Specifically, as an optional embodiment, the method also includes: when the server determines that a violation has occurred on the road based on the analysis results, it generates time period information and event location, and sends them to the drone; the drone receives the time period information, obtains the first video data shot by the first shooting component, and intercepts the target video data based on the time period information, and uploads it to the server; the drone adjusts its position according to the event location, and shoots the second video data based on the first shooting component, and uploads it to the server.

[0025] In this solution, the drone can shoot images after adjusting the focal length according to the focal length information, or can use the autofocus function to perform autofocus, and finally compare the clarity of the two images to select the clearer image as the second image. Specifically, as an optional embodiment, the drone controls the second shooting component to adjust the focal length according to the focal length information at the shooting position, and obtains the second image of each target object in the first image, including: the drone controls the second shooting component to adjust the focal length according to the focal length information at the shooting position, and obtains the first intermediate image; the drone controls the second shooting component at the shooting position to focus on the target object and obtain the second intermediate image; and compares the clarity of the first intermediate image and the second intermediate image to obtain the second image. The image taken according to the focal length information may not be clear, and the image quality can be improved by using autofocus.

[0026] The drone can generate a focal range according to the focal length information, then take values ​​multiple times within the focal range and then capture images, and compare the clarity between the images to obtain the clearest image. Specifically, as an optional embodiment, the drone controls the second shooting component at the shooting position to focus the target object and obtain a second intermediate image, including: the drone runs to the shooting position and determines the focal range according to the focal length information; obtains multiple captured images at the capture interval within the focal range, and compares the clarity of the multiple captured images, so as to reduce the focal range and the capture interval according to the focal length of the clearest captured image, so as to obtain the clearest captured image as the second intermediate image, and the step of comparing the image clarity is determined by comparing the brightness values ​​of the first pixel of the target object and the second pixel adjacent to the first pixel in the captured image.

[0027] In addition, the present solution can use two shooting components of the drone to shoot a video of the surrounding environment of the road. Since the focal lengths of the two shooting components are different, one shooting can be regarded as two shots of a single camera, so that the server can use two videos to build a model and obtain a road model. Specifically, as an optional embodiment, the area shot by the first shooting component and the second shooting component of the drone at least partially overlaps, and the method also includes: the drone obtains a first environmental video related to the road through the first shooting component at the environment acquisition position, and obtains a second environmental video related to the road through the second shooting component, the first shooting component shoots the video with the first shooting focal length, and the second shooting component shoots the video with the second shooting focal length; the drone uploads the first video data, the first shooting focal length, the second video data, the second shooting focal length and the environment acquisition position of the drone to the server, so as to build a model on the server to form a road model. After determining the road model, the drone can upload a clear image containing the target object. For the area that does not contain the target object, some pixels can be screened out to obtain pixel values, and the pixel values ​​of other pixels are adjusted accordingly in combination with the road model, so as to restore it to the target image. Specifically, as an optional embodiment, determining the first area of ​​the target object in the first image, replacing the image of the first area according to the second image, determining the target image to be uploaded, and uploading the target image to the server, includes: determining the first area containing the target object and the second area not containing the target object in the first image; cropping the object image containing the target object from the second image, replacing the image of the first area according to the object image, and uploading it to the server; the drone uploads the parameter information when the first shooting component shoots the first image to the server, and extracts the pixel values ​​of multiple fixed pixels from the second area of ​​the first image, and uploads it to the server, the server extracts the background image from the environment model according to the parameter information, and renders the background image according to the pixel values ​​of multiple fixed pixels, and restores the image in combination with the object image to obtain the restored target image, the parameter information includes the position of the drone and the orientation of the first shooting component. The server determines the area corresponding to the first image in the road model according to the position and orientation of the drone, and then intercepts the image to generate the background.

[0028] After the server acquires the image, it can determine whether the image contains noise based on the weather. If there may be noise in the image, it can acquire images before and after the target object and analyze the noise based on the correlation to perform denoising, so as to perform image analysis based on the denoised image. Specifically, as an optional embodiment, the server receives the target image, and adopts an analysis model based on a tensor processor to analyze the target object, and the steps of determining the analysis result include: the server receives the target image, and obtains noise-related parameters related to the road, and the noise-related parameters include weather information; when it is determined that the road contains noise based on the noise-related parameters, the server acquires the relevant image of the target image from the drone, and performs denoising on the target image based on the target image and the relevant image to obtain a denoised image, and the relevant image of the target image includes the image before or after the target image; the denoised image is input into the analysis model based on the tensor processor for rapid analysis to determine the analysis result, and the analysis result includes the violation analysis result or the vehicle information analysis result.

[0029] On the basis of the above embodiments, the embodiments of the present application further provide a rapid image recognition system based on a drone, wherein the drone comprises a first shooting component and a second shooting component, wherein the first shooting component is used to obtain a first image of a road, and the second shooting component is used to focus on target objects in the first image respectively to obtain a second image of the target object, such as Figure 2 As shown, the system comprises: The first image acquisition module 202 is used to acquire a first image of the road through a first shooting component of the drone, where the first image contains a target object.

[0030] The position information recognition module 204 is used to recognize the position information of the target object in the first image, and determine the shooting position and focal length information of the drone according to the position information.

[0031] The second image acquisition module 206 is used to acquire a second image of each target object in the first image after the drone controls the second shooting component at the shooting position to adjust the focal length according to the focal length information.

[0032] The target image determination module 208 is used to determine the first area of ​​the target object in the first image, and replace the image of the first area according to the second image, so as to determine the target image to be uploaded, so as to upload the target image to the server. The server receives the target image, analyzes the target object using an analysis model based on a tensor processor, and determines the analysis result.

[0033] The implementation of the embodiment of the present application is similar to the implementation of the embodiment of the method described above. The specific implementation can refer to the specific implementation of the method described above, which will not be repeated here.

[0034] This application can be applied in the scenario of acquiring road images and analyzing the images. This solution can use the movement of the drone to acquire a clearer image of the target object (vehicle) on the road, and upload the clear image to the server for analysis. The drone may include a first shooting component and a second shooting component. The first shooting component is used to acquire a first image of the road, and the second shooting component is used to focus on the target object in the first image respectively to acquire a second image of the target object. The drone can acquire the first image through the first shooting component, and use the long-distance focusing ability of the second shooting component to acquire a clear second image of the target object, and then merge the first image and the second image and upload them to the server. This solution can obtain a clearer vehicle image for analysis. Specifically, the first image of the road can be obtained by the first shooting component of the drone, and the first image contains the target object; then the position information of the target object in the first image is identified, and the shooting position and focal length information of the drone are determined according to the position information; the drone controls the second shooting component to adjust the focal length according to the focal length information at the shooting position, and obtains the second image of each target object in the first image; the first area of ​​the target object in the first image is determined, and the image of the first area is replaced according to the second image, and the target image to be uploaded is determined, so as to upload the target image to the server, and the server receives the target image, and adopts the analysis model based on the tensor processor to analyze the target object and determine the analysis result. The shooting position in this scheme can be the position where the drone shoots the first image, or it can be the position after moving based on the position where the first image is shot. This scheme focuses on the position of the target object, and will not focus on objects that are too close or too far away in the case of noise interference (such as rain, snow, fog, etc.), so that a clearer image can be obtained. This scheme can use the tensor processor to quickly realize image processing and analysis to determine the violation analysis results and vehicle information analysis results.

[0035] Based on the above embodiments, the present application also provides an electronic device, comprising: a memory and at least one processor; the memory is used to store computer execution instructions; the at least one processor is used to execute the computer execution instructions stored in the memory, so that the at least one processor executes the method described in the above embodiments.

[0036] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, each process of the above-mentioned data processing method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0037] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented 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.

[0038] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0039] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0040] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0041] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0042] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0043] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. According to the definition in this article, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0044] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0045] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented 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.

[0046] The above are only embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method for rapid image recognition based on drones, characterized in that: The drone includes a first shooting component and a second shooting component, the first shooting component is used to obtain a first image of a road, and the second shooting component is used to focus on a target object in the first image to obtain a second image of the target object. The method includes: The first shooting component of the drone acquires a first image of the road, where the first image includes the target object; Identify the position information of the target object in the first image, and determine the shooting position and focal length information of the drone according to the position information; The drone controls the second shooting component at the shooting position to adjust the focal length according to the focal length information, and obtains a second image of each target object in the first image; Determine a first area of ​​a target object in a first image, replace the image of the first area according to a second image, determine a target image to be uploaded, upload the target image to a server, and the server receives the target image to analyze the target object using an analysis model based on a tensor processor to determine an analysis result.

2. The method according to claim 1, characterized in that The method further comprises: When the server determines that a traffic violation has occurred on the road based on the analysis results, it generates time period information and event location and sends it to the drone; The drone receives the time period information, obtains the first video data shot by the first shooting component, and intercepts the target video data according to the time period information, and uploads the data to the server; The drone adjusts its position according to the event location, and shoots second video data according to the first shooting component, and uploads the second video data to the server.

3. The method according to claim 2, characterized in that The drone controls the second shooting component to adjust the focal length according to the focal length information at the shooting position, and obtains a second image of each target object in the first image, including: The drone controls the second shooting component at the shooting position to adjust the focal length according to the focal length information, and obtains a first intermediate image; The drone controls the second shooting component at the shooting position to focus the target object and obtain a second intermediate image; The clarity of the first intermediate image and the second intermediate image are compared to obtain a second image.

4. The method according to claim 3, characterized in that The drone controls the second shooting component at the shooting position to focus the target object to obtain a second intermediate image, including: The drone moves to the shooting position and determines the focal length range based on the focal length information; Acquire multiple acquired images within the focal length range according to the acquisition interval, and compare the clarity of the multiple acquired images, so as to reduce the focal length range and the acquisition interval according to the focal length of the clearest acquired image, so as to obtain the clearest acquired image as the second intermediate image. The step of comparing the image clarity is to compare the brightness values ​​of the first pixel point of the target object in the acquired image and the second pixel point adjacent to the first pixel point to determine.

5. The method according to claim 4, characterized in that The areas captured by the first shooting component and the second shooting component of the drone at least partially overlap, and the method further includes: The drone is at an environment collection position, and obtains a first environment video related to the road through a first shooting component, and obtains a second environment video related to the road through a second shooting component. The first shooting component shoots the video at a first shooting focal length, and the second shooting component shoots the video at a second shooting focal length. The drone uploads first video data, a first shooting focal length, second video data, a second shooting focal length, and an environment acquisition position of the drone to a server for modeling at the server to form a road model.

6. The method according to claim 5, characterized in that Determining a first area of ​​a target object in a first image, replacing an image of the first area according to a second image, determining a target image to be uploaded, and uploading the target image to a server, including: Determine a first area including the target object and a second area not including the target object in the first image; Cropping an object image containing the target object from the second image, replacing the image of the first area according to the object image, and uploading it to the server; The drone uploads the parameter information of the first shooting component when shooting the first image to the server, and extracts the pixel values ​​of multiple fixed pixels from the second area of ​​the first image, and uploads them to the server. The server extracts the background image from the environment model according to the parameter information, and renders the background image according to the pixel values ​​of the multiple fixed pixels, and restores the image in combination with the object image to obtain a restored target image. The parameter information includes the position of the drone and the orientation of the first shooting component.

7. The method according to claim 6, characterized in that The server receives the target image, analyzes the target object using the tensor processor-based analysis model, and determines the analysis result, including: The server receives the target image and obtains road-related noise-related parameters, including weather information; When it is determined that the road contains noise according to the noise-related parameters, the server obtains the relevant images of the target image from the drone, and performs denoising processing on the target image according to the target image and the relevant images to obtain the denoised image, wherein the relevant images of the target image include images before or after the target image; The denoised image is input into an analysis model based on a tensor processor for rapid analysis to determine analysis results, which include violation analysis results or vehicle information analysis results.

8. A rapid image recognition system based on drones, characterized in that: The drone includes a first shooting component and a second shooting component, the first shooting component is used to obtain a first image of a road, and the second shooting component is used to focus on a target object in the first image to obtain a second image of the target object. The system includes: A first image acquisition module, used to acquire a first image of the road through a first shooting component of the drone, wherein the first image includes a target object; A position information recognition module, used to recognize the position information of the target object in the first image, and determine the shooting position and focal length information of the drone according to the position information; A second image acquisition module is used to acquire a second image of each target object in the first image after the drone controls the second shooting component to adjust the focal length according to the focal length information at the shooting position; The target image determination module is used to determine the first area of ​​the target object in the first image, and replace the image of the first area according to the second image, so as to determine the target image to be uploaded, so as to upload the target image to the server. The server receives the target image, analyzes the target object using an analysis model based on a tensor processor, and determines the analysis result.

9. An electronic device, characterized in that: include: memory and at least one processor; The memory is used to store computer-executable instructions; The at least one processor is configured to execute computer-executable instructions stored in the memory, so that the at least one processor performs the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored on a computer-readable storage medium, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images