Unmanned aerial vehicle control method and device, computer equipment, storage medium and computer program product

Abstract

The invention relates to an unmanned aerial vehicle control method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring a to-be-analyzed video which is shot by a to-be-controlled unmanned aerial vehicle and is associated with distribution network equipment; performing framing processing on the to-be-analyzed video to obtain a plurality of to-be-analyzed video frames; performing conversion processing on each to-be-analyzed video frame to obtain text information corresponding to each to-be-analyzed video frame; obtaining a word vector of the text information corresponding to each to-be-analyzed video frame; performing fusion processing on the word vector of the text information corresponding to each to-be-analyzed video frame to obtain a fused word vector; inputting the fusion word vector into a trained unmanned aerial vehicle control instruction prediction model to obtain a target unmanned aerial vehicle control instruction corresponding to the fusion word vector; and controlling the to-be-controlled unmanned aerial vehicle to monitor the distribution network equipment according to the target unmanned aerial vehicle control instruction. By adopting the method, the control efficiency of the unmanned aerial vehicle can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method and apparatus for controlling an unmanned aerial vehicle, a computer device, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the development of drone technology, using drones to efficiently handle tasks is crucial to improving task processing efficiency.

[0003] In traditional technology, manual control is generally used to control drones; however, this manual control method easily consumes a lot of time and manpower, resulting in low control efficiency of the drone. Summary of the invention

[0004] Based on this, it is necessary to provide a drone control method, device, computer equipment, computer-readable storage medium and computer program product that can improve the control efficiency of the drone in response to the above technical problems.

[0005] In a first aspect, the present application provides a method for controlling a drone, comprising:

[0006] Obtain the video to be analyzed that is taken by the drone to be controlled and is associated with the network distribution device;

[0007] Performing frame processing on the video to be analyzed to obtain multiple video frames to be analyzed;

[0008] Performing conversion processing on each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed;

[0009] Obtain the word vector of the text information corresponding to each video frame to be analyzed;

[0010] Performing fusion processing on the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector;

[0011] Inputting the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector;

[0012] According to the target UAV control instruction, the UAV to be controlled is controlled to monitor the distribution network equipment.

[0013] In one embodiment, the step of performing frame processing on the video to be analyzed to obtain a plurality of video frames to be analyzed includes:

[0014] Performing frame processing on the video to be analyzed to obtain multiple candidate video frames;

[0015] Perform denoising on each candidate video frame to obtain multiple processed video frames;

[0016] The plurality of processed video frames are extracted and processed at preset time intervals to obtain the plurality of video frames to be analyzed.

[0017] In one embodiment, the fusing process of the word vectors of the text information corresponding to each of the video frames to be analyzed to obtain the fused word vectors includes:

[0018] Performing word segmentation processing on the text information corresponding to each of the video frames to be analyzed to obtain multiple word segments of the text information corresponding to each of the video frames to be analyzed;

[0019] Determining the importance of the text information corresponding to each of the video frames to be analyzed according to the information entropy of multiple word segments of the text information corresponding to each of the video frames to be analyzed;

[0020] Determining the weight of the text information corresponding to each video frame to be analyzed according to the importance of the text information corresponding to each video frame to be analyzed;

[0021] According to the weight of the text information corresponding to each video frame to be analyzed, the word vectors of the text information corresponding to each video frame to be analyzed are fused to obtain a fused word vector.

[0022] In one embodiment, inputting the fused word vector into a trained UAV control instruction prediction model to obtain a target UAV control instruction corresponding to the fused word vector includes:

[0023] Inputting the fused word vector into the trained UAV control instruction prediction model to obtain a plurality of preset UAV control instructions corresponding to the fused word vector and a prediction probability corresponding to each of the preset UAV control instructions;

[0024] From the preset UAV control instructions, the corresponding preset UAV control instruction with the largest prediction probability is selected as the target UAV control instruction.

[0025] In one embodiment, the trained UAV control instruction prediction model is trained in the following manner:

[0026] Obtain a sample video associated with a sample network distribution device shot by a sample drone;

[0027] Performing frame processing on the sample video to obtain multiple sample video frames;

[0028] Performing conversion processing on each sample video frame to obtain text information corresponding to each sample video frame;

[0029] Obtain a word vector of the text information corresponding to each sample video frame;

[0030] Performing fusion processing on the word vectors of the text information corresponding to each sample video frame to obtain a sample fused word vector;

[0031] Inputting the sample fusion word vector into the unmanned aerial vehicle control instruction prediction model to be trained to obtain the predicted unmanned aerial vehicle control instruction corresponding to the sample fusion word vector;

[0032] The actual UAV control instruction corresponding to the sample fusion word vector is obtained, and according to the difference between the predicted UAV control instruction and the actual UAV control instruction, the UAV control instruction prediction model to be trained is iteratively trained to obtain the trained UAV control instruction prediction model.

[0033] In one embodiment, before obtaining the video to be analyzed that is taken by the drone to be controlled and is associated with the network distribution device, the method further includes:

[0034] Obtaining first location information of a distribution network device and second location information of multiple candidate drones;

[0035] Determine the distance information between each of the candidate drones and the network distribution device according to the first location information of the network distribution device and the second location information of multiple candidate drones;

[0036] From the candidate drones, a candidate drone with the smallest distance information to the network distribution device is selected as the drone to be controlled.

[0037] In a second aspect, the present application also provides a drone control device, comprising:

[0038] A video acquisition module is used to acquire the video to be analyzed that is taken by the drone to be controlled and is associated with the distribution network device;

[0039] A video processing module, used for performing frame processing on the video to be analyzed to obtain multiple video frames to be analyzed;

[0040] The video frame conversion module is used to convert each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed;

[0041] A word vector acquisition module, used to obtain the word vector of the text information corresponding to each video frame to be analyzed;

[0042] A word vector fusion module is used to fuse the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector;

[0043] An instruction prediction module, used to input the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector;

[0044] The drone control module is used to control the drone to be controlled to monitor the distribution network equipment according to the target drone control instruction.

[0045] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0046] Obtain the video to be analyzed that is taken by the drone to be controlled and is associated with the network distribution device;

[0047] Performing frame processing on the video to be analyzed to obtain multiple video frames to be analyzed;

[0048] Performing conversion processing on each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed;

[0049] Obtain the word vector of the text information corresponding to each video frame to be analyzed;

[0050] Performing fusion processing on the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector;

[0051] Inputting the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector;

[0052] According to the target UAV control instruction, the UAV to be controlled is controlled to monitor the distribution network equipment.

[0053] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0054] Obtain the video to be analyzed that is taken by the drone to be controlled and is associated with the network distribution device;

[0055] Performing frame processing on the video to be analyzed to obtain multiple video frames to be analyzed;

[0056] Performing conversion processing on each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed;

[0057] Obtain the word vector of the text information corresponding to each video frame to be analyzed;

[0058] Performing fusion processing on the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector;

[0059] Inputting the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector;

[0060] According to the target UAV control instruction, the UAV to be controlled is controlled to monitor the distribution network equipment.

[0061] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0062] Obtain the video to be analyzed that is taken by the drone to be controlled and is associated with the network distribution device;

[0063] Performing frame processing on the video to be analyzed to obtain multiple video frames to be analyzed;

[0064] Performing conversion processing on each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed;

[0065] Obtain the word vector of the text information corresponding to each video frame to be analyzed;

[0066] Performing fusion processing on the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector;

[0067] Inputting the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector;

[0068] According to the target UAV control instruction, the UAV to be controlled is controlled to monitor the distribution network equipment.

[0069] The above-mentioned drone control method, device, computer equipment, storage medium and computer program product first obtain the video to be analyzed associated with the distribution network equipment shot by the drone to be controlled, and perform frame processing on the video to be analyzed to obtain multiple video frames to be analyzed, and then convert each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed, then obtain the word vector of the text information corresponding to each video frame to be analyzed, and fuse the word vector of the text information corresponding to each video frame to be analyzed to obtain a fused word vector, then input the fused word vector into the trained drone control instruction prediction model to obtain the target drone control instruction corresponding to the fused word vector, and finally, according to the target drone control instruction, control the drone to be controlled to monitor the distribution network equipment. In this way, in the process of controlling the UAV, through frame processing and text conversion, the video information can be quickly converted into analyzable text form, reducing the complexity of directly processing video data; secondly, through word vector fusion and the use of trained models to predict control instructions, the speed of instruction generation is greatly improved, which is conducive to improving the control efficiency of the UAV; moreover, the entire process does not require human intervention, avoiding the defect of manual control that is easy to consume a lot of time and manpower, resulting in low control efficiency of the UAV, and further improving the control efficiency of the UAV. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0071] Figure 1 A schematic diagram of a flow chart of a method for controlling a drone in one embodiment;

[0072] Figure 2 A schematic diagram of a flow chart of a drone control method in another embodiment;

[0073] Figure 3 A schematic diagram of a low-altitude UAV remote control system in one embodiment;

[0074] Figure 4 A schematic diagram of a low-altitude UAV remote control system in another embodiment;

[0075] Figure 5 is a structural block diagram of a drone control device in one embodiment;

[0076] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0077] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0078] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the flight altitude and flight range of the drone, and the collection, use and processing of related data must comply with relevant regulations.

[0079] In an exemplary embodiment, Figure 1 As shown, a method for controlling a drone is provided. This embodiment uses the method applied to a server as an example for illustration. It is understandable that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smart phones, and tablet computers; the server can be implemented as an independent server or a server cluster consisting of multiple servers. In this embodiment, the method includes the following steps:

[0080] Step S101, obtaining a video to be analyzed that is taken by the drone to be controlled and is associated with a network distribution device.

[0081] The drone to be controlled refers to a drone with a shooting function that needs to be controlled.

[0082] Among them, distribution network equipment refers to various equipment in the power distribution network, such as transformers, switch cabinets, distribution boxes, etc.

[0083] The video to be analyzed refers to the video shot by the drone to be controlled that contains the network distribution equipment.

[0084] Exemplarily, in response to a monitoring instruction for a distribution network device, the server determines a drone used to photograph the distribution network device as the drone to be controlled; then, the server filters out videos containing the distribution network device from the videos taken by the drone to be controlled as the videos to be analyzed that are taken by the drone to be controlled and are associated with the distribution network device.

[0085] Step S102, performing frame processing on the video to be analyzed to obtain a plurality of video frames to be analyzed.

[0086] The frame processing refers to a process of obtaining a plurality of video frames to be analyzed based on the video to be analyzed.

[0087] The video frame to be analyzed refers to the video frame that needs to be analyzed.

[0088] Exemplarily, the server preprocesses the video to be analyzed to obtain a preprocessed video; for example, the server converts the format of the video to be analyzed to obtain a preprocessed video; then, the server performs frame processing on the preprocessed video to obtain multiple video frames to be analyzed.

[0089] Step S103: convert each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed.

[0090] Among them, text information refers to the content described in words.

[0091] Exemplarily, the server inputs each video frame to be analyzed into a text prediction model (such as a convolutional neural network), converts each video frame to be analyzed through the text prediction model, and obtains text information corresponding to each video frame to be analyzed.

[0092] Step S104, obtaining the word vector of the text information corresponding to each video frame to be analyzed.

[0093] Among them, the word vector is used to represent the representation vector of the text information corresponding to each video frame to be analyzed.

[0094] Exemplarily, the server preprocesses the text information corresponding to each video frame to be analyzed to obtain preprocessed text information; then, the server inputs the preprocessed text information into a word vector model, converts the preprocessed text information through the word vector model, and obtains the word vector of the text information corresponding to each video frame to be analyzed.

[0095] Step S105, fusing the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector.

[0096] The fusion processing may be a weighted summation processing.

[0097] The fused word vector refers to the word vector obtained by fusing the word vectors of the text information corresponding to each video frame to be analyzed.

[0098] Exemplarily, the server determines the weight of the word vector of the text information corresponding to each video frame to be analyzed; then, the server fuses the word vector of the text information corresponding to each video frame to be analyzed according to the weight of the word vector of the text information corresponding to each video frame to be analyzed to obtain a fused word vector; for example, the server performs weighted summation processing on the word vector of the text information corresponding to each video frame to be analyzed according to the weight of the word vector of the text information corresponding to each video frame to be analyzed to obtain a fused word vector.

[0099] Step S106, input the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector.

[0100] Among them, the UAV control instruction prediction model refers to a network model that can use the fused word vector to obtain the target UAV control instruction corresponding to the fused word vector.

[0101] Among them, the target UAV control instruction refers to the instruction information corresponding to the control of the UAV to be controlled.

[0102] Exemplarily, the server inputs the fused word vector into a trained UAV control instruction prediction model, and obtains the target UAV control instruction corresponding to the fused word vector through the UAV control instruction prediction model.

[0103] Step S107, controlling the drone to be controlled to monitor the distribution network equipment according to the target drone control instruction.

[0104] Exemplarily, the server parses and processes the control instructions of the target drone to obtain parsed information; then, the server determines the drone operation information (such as flight altitude, flight speed, shooting angle, and monitoring area range) in the parsed information, and controls the drone to be controlled to monitor the distribution network equipment based on the drone operation information.

[0105] In the above-mentioned UAV control method, firstly, the video to be analyzed and associated with the distribution network equipment shot by the UAV to be controlled is obtained, and the video to be analyzed is processed by frame division to obtain multiple video frames to be analyzed, and then each video frame to be analyzed is converted to obtain the text information corresponding to each video frame to be analyzed, and then the word vector of the text information corresponding to each video frame to be analyzed is obtained, and the word vector of the text information corresponding to each video frame to be analyzed is fused to obtain the fused word vector, and then the fused word vector is input into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector, and finally, according to the target UAV control instruction, the UAV to be controlled is controlled to monitor the distribution network equipment. In this way, in the process of controlling the UAV, through frame division processing and text conversion, the video information can be quickly converted into an analyzable text form, which reduces the complexity of directly processing video data; secondly, through word vector fusion and using the trained model to predict the control instruction, the speed of instruction generation is greatly improved, which is conducive to improving the control efficiency of the UAV; moreover, the whole process does not require manual intervention, avoiding the defect that the manual control method is easy to consume a lot of time and manpower, resulting in low control efficiency of the UAV, and further improving the control efficiency of the UAV.

[0106] In an exemplary embodiment, the above step S102, in which the video to be analyzed is divided into frames to obtain a plurality of video frames to be analyzed, specifically includes the following contents: the video to be analyzed is divided into frames to obtain a plurality of candidate video frames; each candidate video frame is denoised to obtain a plurality of processed video frames; and the plurality of processed video frames are extracted at preset time intervals to obtain a plurality of video frames to be analyzed.

[0107] The candidate video frames refer to all video frames obtained by performing frame segmentation processing on the video to be analyzed.

[0108] The processed video frame refers to the candidate video frame after denoising.

[0109] The preset time interval refers to a pre-set time interval, such as 10 seconds.

[0110] Exemplarily, the server performs frame processing on the video to be analyzed according to a preset frame processing method to obtain multiple candidate video frames; for example, the server performs frame processing on the video to be analyzed at an interval of one second to obtain video frames corresponding to each second as multiple candidate video frames; then, the server performs denoising processing on each candidate video frame to obtain multiple processed video frames; then, the server performs extraction processing on the multiple processed video frames at a preset time interval to obtain multiple video frames to be analyzed; for example, the server performs extraction processing on the multiple processed video frames at a time interval of every 10 seconds to obtain video frames at the 1st second, the 11th second, the 21st second, etc. as multiple video frames to be analyzed.

[0111] In this embodiment, through the process of denoising and extracting key frames, noise interference and unnecessary frames in the video to be analyzed can be reduced, thereby reducing the consumption of computing resources, which is beneficial to improving the processing speed of the video to be analyzed and further improving the extraction efficiency of the video frame to be analyzed.

[0112] In an exemplary embodiment, the above-mentioned step S105, fusing the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector, specifically includes the following contents: performing word segmentation processing on the text information corresponding to each video frame to be analyzed to obtain multiple word segments of the text information corresponding to each video frame to be analyzed; determining the importance of the text information corresponding to each video frame to be analyzed based on the information entropy of the multiple word segments of the text information corresponding to each video frame to be analyzed; determining the weight of the text information corresponding to each video frame to be analyzed based on the importance of the text information corresponding to each video frame to be analyzed; and fusing the word vectors of the text information corresponding to each video frame to be analyzed based on the weight of the text information corresponding to each video frame to be analyzed to obtain a fused word vector.

[0113] The word segmentation process refers to the process of obtaining multiple word segments corresponding to the text information based on the text information.

[0114] Among them, word segmentation refers to the words in the text information.

[0115] Among them, information entropy is used to characterize the uncertainty of text information.

[0116] Among them, importance refers to the importance of text information.

[0117] Among them, the weight is used to represent the importance of text information.

[0118] Exemplarily, the server performs word segmentation processing on the text information corresponding to each video frame to be analyzed according to a preset word segmentation method (such as a forward maximum matching method) to obtain multiple word segments of the text information corresponding to each video frame to be analyzed; then, the server determines the information entropy of the multiple word segments of the text information corresponding to each video frame to be analyzed according to the occurrence frequency of the multiple word segments of the text information corresponding to each video frame to be analyzed; then, the server determines the importance of the information entropy of the multiple word segments of the text information corresponding to each video frame to be analyzed according to the information entropy of the multiple word segments of the text information corresponding to each video frame to be analyzed, as the importance of the text information corresponding to each video frame to be analyzed; then, the server sorts the text information corresponding to each video frame to be analyzed according to the importance of the text information corresponding to each video frame to be analyzed, and obtains the sorting information of the text information corresponding to each video frame to be analyzed; then, the server determines the weight of the text information corresponding to each video frame to be analyzed according to the sorting information of the text information corresponding to each video frame to be analyzed; then, the server performs fusion processing on the word vectors of the text information corresponding to each video frame to be analyzed according to the weight of the text information corresponding to each video frame to be analyzed, and obtains the fused word vector.

[0119] In this embodiment, by determining the weight of the text information corresponding to each video frame to be analyzed, and based on the weight, the word vectors of the text information corresponding to each video frame to be analyzed are fused, so that text information with higher importance is given higher weights, which will have a greater impact on the final fused word vector during the fusion process, and is conducive to obtaining a more accurate fused word vector, thereby improving the accuracy of determining the fused word vector.

[0120] In an exemplary embodiment, the above step S106, inputting the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector, specifically includes the following contents: inputting the fused word vector into the trained UAV control instruction prediction model to obtain multiple preset UAV control instructions corresponding to the fused word vector, and the prediction probability corresponding to each preset UAV control instruction; from each preset UAV control instruction, screening out the preset UAV control instruction with the largest corresponding prediction probability as the target UAV control instruction.

[0121] Among them, the preset drone control instructions refer to pre-set drone control instructions, such as rise, fall, forward, backward, move left, and move right.

[0122] Among them, the prediction probability is used to characterize the possibility that the UAV control instruction prediction model determines that the preset UAV control instruction is correct.

[0123] Exemplarily, the server inputs the fused word vector into the trained UAV control instruction prediction model to obtain multiple preset UAV control instructions corresponding to the fused word vector and the prediction probability corresponding to each preset UAV control instruction; then, the server sorts each preset UAV control instruction according to the prediction probability corresponding to each preset UAV control instruction to obtain the sorted preset UAV control instructions; then, the server selects the preset UAV control instruction with the largest corresponding prediction probability from the sorted preset UAV control instructions as the target UAV control instruction.

[0124] In this embodiment, by screening out the preset drone control instructions with the highest corresponding predicted probability as the target drone control instructions, the defect of low drone control accuracy caused by uncertainty in randomly selecting instructions is avoided, thereby improving the drone control accuracy.

[0125] In an exemplary embodiment, the drone control method provided in the present application also includes a training step for a trained drone control instruction prediction model, which specifically includes the following contents: obtaining a sample video associated with a sample distribution network device shot by a sample drone; performing frame processing on the sample video to obtain multiple sample video frames; performing conversion processing on each sample video frame to obtain text information corresponding to each sample video frame; obtaining a word vector of the text information corresponding to each sample video frame; performing fusion processing on the word vector of the text information corresponding to each sample video frame to obtain a sample fusion word vector; inputting the sample fusion word vector into the drone control instruction prediction model to be trained to obtain a predicted drone control instruction corresponding to the sample fusion word vector; obtaining the actual drone control instruction corresponding to the sample fusion word vector, and iteratively training the drone control instruction prediction model to be trained based on the difference between the predicted drone control instruction and the actual drone control instruction to obtain a trained drone control instruction prediction model.

[0126] Among them, the sample drone refers to the drone corresponding to the drone control command prediction model used for training.

[0127] Among them, the network distribution equipment corresponding to the UAV control command prediction model is used to train the UAV control command prediction model.

[0128] The sample video refers to the video shot by the sample drone and containing the sample distribution network equipment.

[0129] Among them, the sample video frames refer to the video frames corresponding to the UAV control command prediction model used to train.

[0130] The sample fused word vector refers to the word vector obtained by fusing the word vectors of the text information corresponding to each sample video frame.

[0131] Among them, the predicted UAV control instruction refers to the predicted value of the UAV control instruction corresponding to the sample fusion word vector.

[0132] Among them, the actual UAV control instruction refers to the actual value of the UAV control instruction corresponding to the sample fusion word vector.

[0133] Exemplarily, in response to a model training instruction for a drone control instruction prediction model, the server obtains a sample video associated with a sample distribution network device shot by a sample drone from a database; then, the server performs frame processing on the sample video to obtain a plurality of sample video frames; then, the server performs conversion processing on each sample video frame to obtain text information corresponding to each sample video frame; then, the server obtains a word vector of the text information corresponding to each sample video frame; then, the server performs fusion processing on the word vector of the text information corresponding to each sample video frame to obtain a sample fusion word vector; then, the server inputs the sample fusion word vector into the drone control instruction prediction model to be trained. The predicted UAV control instruction corresponding to the sample fusion word vector is obtained; then, the server obtains the actual UAV control instruction corresponding to the sample fusion word vector, and obtains the loss value according to the difference between the predicted UAV control instruction and the actual UAV control instruction; then, the server adjusts the model parameters of the UAV control instruction prediction model to be trained according to the loss value; then, the server re-trains the UAV control instruction prediction model after the model parameters are adjusted until the loss value obtained by the trained UAV control instruction prediction model is less than the loss value threshold, then the training is stopped, and the trained UAV control instruction prediction model is used as the trained UAV control instruction prediction model.

[0134] In this embodiment, by pre-training the UAV control command prediction model, it is convenient to predict the target UAV control command corresponding to the fused word vector after obtaining the fused word vector in actual application; moreover, the UAV control command prediction model receives new data in each round of iteration, and performs internal model improvements and optimizations to facilitate more effective predictions, which is beneficial to improving the prediction accuracy of the UAV control command prediction model.

[0135] In an exemplary embodiment, the above-mentioned step S101, before obtaining the video to be analyzed associated with the network distribution device shot by the drone to be controlled, specifically includes the following contents: obtaining the first position information of the network distribution device and the second position information of multiple candidate drones; determining the distance information between each candidate drone and the network distribution device based on the first position information of the network distribution device and the second position information of multiple candidate drones; and screening out the candidate drone with the smallest distance information to the network distribution device from each candidate drone as the drone to be controlled.

[0136] The first location information is used to represent the three-dimensional coordinate information of the distribution network equipment.

[0137] Among them, the candidate drone refers to the drone to be selected.

[0138] The second position information is used to represent the three-dimensional coordinate information of the candidate drone.

[0139] The distance information may refer to the horizontal distance between each candidate UAV and the distribution network equipment.

[0140] Exemplarily, the server obtains the three-dimensional coordinate information of the network distribution device through the global positioning system as the first position information of the network distribution device; then, the server obtains the three-dimensional coordinate information of the candidate drones through the global navigation satellite system as the second position information of multiple candidate drones; then, the server determines the horizontal distance between each candidate drone and the network distribution device based on the first position information of the network distribution device and the second position information of multiple candidate drones, which corresponds to the distance information between each candidate drone and the network distribution device; then, the server selects the candidate drone with the smallest distance information to the network distribution device from each candidate drone as the drone to be controlled.

[0141] In this embodiment, by determining the distance information between each candidate drone and the network distribution device, the drone closest to the network distribution device can be quickly screened out as the drone to be controlled, thereby greatly reducing the time required for the drone to reach the network distribution device, which is beneficial to improving the generation efficiency of the video to be analyzed associated with the network distribution device.

[0142] In an exemplary embodiment, Figure 2 As shown, another drone control method is provided, which is described by taking the method applied to a server as an example, and includes the following steps:

[0143] Step S201, obtaining the first location information of the network distribution device and the second location information of multiple candidate drones; determining the distance information between each candidate drone and the network distribution device according to the first location information of the network distribution device and the second location information of multiple candidate drones.

[0144] Step S202: Select the candidate drone with the smallest distance information to the network distribution device from the candidate drones as the drone to be controlled.

[0145] Step S203, performing frame processing on the video to be analyzed to obtain multiple candidate video frames; performing denoising processing on each candidate video frame to obtain multiple processed video frames; performing extraction processing on the multiple processed video frames at a preset time interval to obtain multiple video frames to be analyzed.

[0146] Step S204: convert each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed.

[0147] Step S205, obtaining the word vector of the text information corresponding to each video frame to be analyzed.

[0148] Step S206, performing word segmentation processing on the text information corresponding to each video frame to be analyzed to obtain multiple word segments of the text information corresponding to each video frame to be analyzed; and determining the importance of the text information corresponding to each video frame to be analyzed based on the information entropy of the multiple word segments of the text information corresponding to each video frame to be analyzed.

[0149] Step S207, determining the weight of the text information corresponding to each video frame to be analyzed according to the importance of the text information corresponding to each video frame to be analyzed; and fusing the word vectors of the text information corresponding to each video frame to be analyzed according to the weight of the text information corresponding to each video frame to be analyzed to obtain a fused word vector.

[0150] Step S208, input the fused word vector into the trained UAV control instruction prediction model to obtain multiple preset UAV control instructions corresponding to the fused word vector and the prediction probability corresponding to each preset UAV control instruction.

[0151] Step S209 , selecting the preset drone control instruction with the highest corresponding prediction probability from the preset drone control instructions as the target drone control instruction.

[0152] Step S210: According to the target UAV control instruction, the UAV to be controlled is controlled to monitor the distribution network equipment.

[0153] In the above-mentioned UAV control method, in the process of controlling the UAV, through frame processing and text conversion, the video information can be quickly converted into analyzable text form, reducing the complexity of directly processing video data; secondly, through word vector fusion and the use of trained models to predict control instructions, the speed of instruction generation is greatly improved, which is conducive to improving the control efficiency of the UAV; moreover, the entire process does not require human intervention, avoiding the defect of manual control that easily consumes a lot of time and manpower, resulting in low control efficiency of the UAV, and further improving the control efficiency of the UAV.

[0154] In an exemplary embodiment, in order to more clearly illustrate the drone control method provided in the embodiment of the present application, the drone control method is specifically described below using a specific embodiment. Figure 3 and Figure 4 As shown, the present application also provides a wireless transmission low-altitude UAV remote control system. In the process of controlling the UAV, the video to be analyzed associated with the distribution network equipment shot by the UAV to be controlled is obtained, and the target UAV control command is generated through the wireless transmission low-altitude UAV remote control system. According to the target UAV control command, the UAV to be controlled is controlled to monitor the distribution network equipment. Specifically, it includes the following contents:

[0155] like Figure 3 As shown, the system includes: data source layer, data integration management layer, business application layer, and customer interaction layer.

[0156] Among them, the data source layer includes relevant data such as cameras, drones, and machine nests. It is the application basis for data management and service interaction. The data in the data source layer is the result of business processing of various business application systems.

[0157] Among them, the data integration management layer mainly realizes the terminal data of cameras, drones, machine nests and other equipment extracted from the data source layer, establishes a physical model through the Internet of Things platform, transmits equipment instructions, controls equipment, realizes map services and route planning, and the video main station accesses the live video stream data. Through the algorithm AI (Artificial Intelligence) provided by the artificial intelligence platform, it analyzes and forms alarm violation data, and realizes the system single sign-on authentication problem by integrating the 4A (Authentication, Authorization, Account, Audit) platform.

[0158] Among them, the business application layer: mainly realizes the customized monitoring and management of the platform's related business functions, including low-altitude remote sensing situation, intelligent operation, achievement management, alarm management, equipment management, flight service, operation and maintenance management, system management, etc. The business application layer follows the big data technical specifications.

[0159] Among them, the customer interaction layer: mainly meets the needs of users to interact with each other during the use of the platform, and realize online recording and service processes, and follows the big data technical specifications.

[0160] like Figure 4 As shown in the figure, it is the internal implementation level of the system. The presentation layer is based on the HTML (HyperText Markup Language) / CSS (Cascading Style Sheets) / JavaScript (a programming language) technology system, and adopts JSP (JavaServer Pages, a dynamic web page technology standard) and Velocity (a template engine based on) template technology to implement RIA (Rich Internet Applications) client. For the lean and practical charts and business process diagrams of the asset management system, Flex (fast lexical analyzer generator) technology is adopted to realize rich graphic presentation.

[0161] The process layer implements the definition and flow of the internal business process of the system based on the BPMN2 (Business Process Model and Notation 2.0) standard. The access layer uses the Apache framework to provide a unified Web Service (network service) interface. The service layer and component layer use Java (a programming language) language, Spring (an application development framework) framework to implement business services and public services, Hibernate (an application development framework) to implement object-relational mapping, and JDBC (Java Database Connectivity) to establish communication with the backend database.

[0162] The system integrates functions including the following technologies:

[0163] 1. Restful (Representational State Transfer) technology. Restful is a technology that can achieve front-end and back-end separation. It directly transmits data by returning a JSON (JavaScript Object Notation, an object representation) data instead of specifying a page corresponding to a string.

[0164] 2. Streaming media technology. Streaming media technology is a new media transmission method that uses streaming transmission technology to continuously play audio, video or multimedia files on the Internet in real time. In streaming media technology, the entire multimedia file is divided into compressed packages through a special compression method, and is continuously and real-timely transmitted from the video server to the user's computer. Users do not have to wait until the entire file is downloaded as in the download method. Instead, they can use the decompression device on the user's computer to decompress the compressed multimedia file and play and watch it after a startup delay of a few seconds or tens of seconds. This streaming transmission method of downloading and playing multimedia files not only greatly shortens the startup delay, but also greatly reduces the demand for system cache capacity.

[0165] 3. EMQX (EMQ eXtensible Messaging Platform) technology. EMQX is an open source IoT MQTT (Message Queuing Telemetry Transport) message agent. It is written in C language and can run on multiple operating systems, including Linux (an operating system), Unix (an operating system), Windows (an operating system), etc. EMQX supports the MQTT protocol, can communicate with various devices that support the MQTT protocol, and provides rich message routing and security functions.

[0166] 4. OAuth 2.0 (Open Authorization 2.0) technology. OAuth 2.0 is an open authorization standard that allows users to authorize third-party applications to access their account information without providing their passwords. It is a common authentication method for many modern APIs (Application Programming Interfaces) and websites. The authentication process of OAuth 2.0 mainly includes the following steps: (1) Request a token: The client requests a token from the authorization server. This token is a temporary credential for accessing user account information. (2) Obtain an access token: The client submits the requested token to the authorization server to obtain permission to access the user account. (3) Access protected resources: The client uses the access token to access protected resources (such as the user's email, photos, etc.). (4) Refresh token: The access token is usually time-limited and needs to be refreshed after expiration. The client can request a new access token from the authorization server.

[0167] It should be noted that in OAuth 2.0, the client can be any application or website that can access the user account. The authorization server is responsible for managing the user's account information and the permissions granted to the client, while the protected resource is the user account itself. OAuth 2.0 has multiple authorization types, including authorization code, implicit authorization, and resource owner password. Among them, the authorization code is the most commonly used method, which sets an intermediate step between the client and the authorization server to ensure that the user's password is not exposed to the client.

[0168] The system has the following interface functions:

[0169] 1. Login unified identity authentication. Through integration with the digital identity and access platform (4A platform), detailed information on the organization and personnel can be obtained to identify the person in charge and work group members of the operation.

[0170] 2. Map service.

[0171] 3. Main distribution network line. Access distribution network line and main network line data.

[0172] 4. Terminal device access. Carry out docking with the Internet of Things platform; based on the four major modules of connection management, device management, application enablement, and operation support of the Internet of Things platform, through open interfaces and protocol plug-ins.

[0173] 5. Unstructured data (document files, audio and video files). By integrating with the intranet and extranet interaction platforms, the interaction between device service data and intranet service data is opened up.

[0174] 6. Streaming media device information. Device information acquisition: The interface can provide detailed information about the device, such as device model, device status, device function, etc.

[0175] 7. Streaming video live broadcast. The interface can receive a variety of video source inputs, such as cameras, drones, etc. The interface supports real-time transmission of video streams and can transmit video streams to a specified server or player to achieve the function of live video broadcast.

[0176] 8. Streaming media device control. Device control command sending, the interface can send control commands to the device through the network or local connection.

[0177] 9. Route management of drone nests. Route planning and design: The interface can provide route planning and design functions according to user needs and flight missions to monitor the route of the drone in real time and provide control functions for the drone.

[0178] 10. Flight control of the drone. Flight control command sending: The interface can send flight control commands to the drone through the network or local connection. The interface can monitor the flight status of the drone in real time, such as position, speed, altitude, battery power, etc.

[0179] 11. Airport information (airport cover). The interface can obtain the status information of the drone airport, including the operation status of the airport, the take-off and landing of the drone, the battery charging status, etc. It can receive and manage the flight plan and mission information of the drone.

[0180] 12. UAV information of the drone nest. The function of the drone information interface of the drone nest is mainly to realize the acquisition, transmission and processing of drone information, including the acquisition of drone status information, the sending of drone control commands, and mission planning and scheduling.

[0181] 13. The artificial intelligence platform network distribution equipment is damaged. The function of the artificial intelligence platform network distribution equipment damage interface is mainly to detect the damage of the network distribution equipment.

[0182] 14. Artificial intelligence platform equipment failure. The function of the artificial intelligence platform distribution network equipment failure interface is mainly to detect distribution network equipment failures.

[0183] 15. Artificial intelligence platform channel hidden dangers. The function of the artificial intelligence platform distribution network channel hidden danger interface is mainly to realize the detection of distribution network channel hidden dangers.

[0184] In the above embodiment, during the process of controlling the drone, the low-altitude drone remote control system with wireless transmission directly obtains the target drone control command according to the video to be analyzed associated with the distribution network equipment shot by the drone to be controlled, so as to control the drone to be controlled to monitor the distribution network equipment. The whole process does not require manual intervention, avoiding the defect that the manual control method is easy to consume a lot of time and manpower, resulting in low control efficiency of the drone, and further improving the control efficiency of the drone. At the same time, different data are obtained by carrying various loads on drones and monitoring devices, and then the ground station transmits the data to the platform, including drone data, machine nest data, positioning data, video data taken and other data (such as routes, flight areas, etc.). The platform integrates the data, including maps, equipment management, personnel information, etc. Commanders log in to the platform at the command center or anywhere else to realize functions such as real-time video monitoring, language intercom, command and dispatch, map review and data display. At the same time, the command center can also use the platform to set route planning to achieve full process management.

[0185] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0186] Based on the same inventive concept, the embodiment of the present application also provides a drone control device for implementing the drone control method involved above. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above method, so the specific limitations in one or more drone control device embodiments provided below can refer to the limitations of the drone control method above, and will not be repeated here.

[0187] In an exemplary embodiment, Figure 5As shown, a drone control device is provided, including: a video acquisition module 501, a video processing module 502, a video frame conversion module 503, a word vector acquisition module 504, a word vector fusion module 505, a command prediction module 506 and a drone control module 507, wherein:

[0188] The video acquisition module 501 is used to acquire the video to be analyzed that is taken by the drone to be controlled and is associated with the network distribution device.

[0189] The video processing module 502 is used to perform frame processing on the video to be analyzed to obtain multiple video frames to be analyzed.

[0190] The video frame conversion module 503 is used to perform conversion processing on each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed.

[0191] The word vector acquisition module 504 is used to obtain the word vector of the text information corresponding to each video frame to be analyzed.

[0192] The word vector fusion module 505 is used to fuse the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector.

[0193] The instruction prediction module 506 is used to input the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector.

[0194] The drone control module 507 is used to control the drone to be controlled to monitor the distribution network equipment according to the target drone control instruction.

[0195] In an exemplary embodiment, the video processing module 502 is also used to perform frame processing on the video to be analyzed to obtain multiple candidate video frames; perform denoising processing on each candidate video frame to obtain multiple processed video frames; and perform extraction processing on the multiple processed video frames at preset time intervals to obtain multiple video frames to be analyzed.

[0196] In an exemplary embodiment, the word vector fusion module 505 is also used to perform word segmentation processing on the text information corresponding to each video frame to be analyzed, so as to obtain multiple word segments of the text information corresponding to each video frame to be analyzed; determine the importance of the text information corresponding to each video frame to be analyzed according to the information entropy of the multiple word segments of the text information corresponding to each video frame to be analyzed; determine the weight of the text information corresponding to each video frame to be analyzed according to the importance of the text information corresponding to each video frame to be analyzed; and perform fusion processing on the word vectors of the text information corresponding to each video frame to be analyzed according to the weight of the text information corresponding to each video frame to be analyzed to obtain a fused word vector.

[0197] In an exemplary embodiment, the instruction prediction module 506 is also used to input the fused word vector into the trained UAV control instruction prediction model to obtain multiple preset UAV control instructions corresponding to the fused word vector, and the prediction probability corresponding to each preset UAV control instruction; from each preset UAV control instruction, the preset UAV control instruction with the largest corresponding prediction probability is screened out as the target UAV control instruction.

[0198] In an exemplary embodiment, the drone control device also includes a model training module, which is used to obtain a sample video associated with a sample distribution network device shot by a sample drone; perform frame processing on the sample video to obtain multiple sample video frames; perform conversion processing on each sample video frame to obtain text information corresponding to each sample video frame; obtain a word vector of the text information corresponding to each sample video frame; perform fusion processing on the word vector of the text information corresponding to each sample video frame to obtain a sample fusion word vector; input the sample fusion word vector into a drone control instruction prediction model to be trained to obtain a predicted drone control instruction corresponding to the sample fusion word vector; obtain the actual drone control instruction corresponding to the sample fusion word vector, and iteratively train the drone control instruction prediction model to be trained based on the difference between the predicted drone control instruction and the actual drone control instruction to obtain a trained drone control instruction prediction model.

[0199] In an exemplary embodiment, the drone control device also includes a drone determination module, which is used to obtain the first position information of the network distribution device and the second position information of multiple candidate drones; determine the distance information between each candidate drone and the network distribution device based on the first position information of the network distribution device and the second position information of multiple candidate drones; from each candidate drone, screen out the candidate drone with the smallest distance information with the network distribution device as the drone to be controlled.

[0200] Each module in the above-mentioned drone control device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0201] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 6As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as videos and text information to be analyzed. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for controlling a drone is implemented.

[0202] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0203] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0204] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0205] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0206] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0207] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0208] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A drone control method, characterized in that: The method comprises: Obtain the video to be analyzed that is taken by the drone to be controlled and is associated with the network distribution device; Performing frame processing on the video to be analyzed to obtain multiple video frames to be analyzed; Performing conversion processing on each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed; Obtain the word vector of the text information corresponding to each video frame to be analyzed; Performing fusion processing on the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector; Inputting the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector; According to the target UAV control instruction, the UAV to be controlled is controlled to monitor the distribution network equipment.

2. The method according to claim 1, characterized in that The step of performing frame processing on the video to be analyzed to obtain a plurality of video frames to be analyzed includes: Performing frame processing on the video to be analyzed to obtain multiple candidate video frames; Perform denoising on each candidate video frame to obtain multiple processed video frames; The plurality of processed video frames are extracted and processed at preset time intervals to obtain the plurality of video frames to be analyzed.

3. The method according to claim 1, characterized in that The fusing the word vectors of the text information corresponding to each of the video frames to be analyzed to obtain a fused word vector includes: Performing word segmentation processing on the text information corresponding to each of the video frames to be analyzed to obtain multiple word segments of the text information corresponding to each of the video frames to be analyzed; Determining the importance of the text information corresponding to each video frame to be analyzed according to the information entropy of multiple word segments of the text information corresponding to each video frame to be analyzed; Determining the weight of the text information corresponding to each video frame to be analyzed according to the importance of the text information corresponding to each video frame to be analyzed; According to the weight of the text information corresponding to each video frame to be analyzed, the word vectors of the text information corresponding to each video frame to be analyzed are fused to obtain a fused word vector.

4. The method according to claim 1, characterized in that: The step of inputting the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector includes: Inputting the fused word vector into the trained UAV control instruction prediction model to obtain a plurality of preset UAV control instructions corresponding to the fused word vector and a prediction probability corresponding to each of the preset UAV control instructions; From the preset UAV control instructions, the corresponding preset UAV control instruction with the largest prediction probability is selected as the target UAV control instruction.

5. The method according to claim 1, characterized in that The trained UAV control command prediction model is trained in the following way: Obtain a sample video associated with a sample network distribution device shot by a sample drone; Performing frame processing on the sample video to obtain multiple sample video frames; Performing conversion processing on each sample video frame to obtain text information corresponding to each sample video frame; Obtain a word vector of the text information corresponding to each sample video frame; Performing fusion processing on the word vectors of the text information corresponding to each sample video frame to obtain a sample fused word vector; Inputting the sample fusion word vector into the unmanned aerial vehicle control instruction prediction model to be trained to obtain the predicted unmanned aerial vehicle control instruction corresponding to the sample fusion word vector; The actual UAV control instruction corresponding to the sample fusion word vector is obtained, and according to the difference between the predicted UAV control instruction and the actual UAV control instruction, the UAV control instruction prediction model to be trained is iteratively trained to obtain the trained UAV control instruction prediction model.

6. The method according to any one of claims 1 to 5, characterized in that Before obtaining the video to be analyzed that is taken by the drone to be controlled and is associated with the network distribution device, the following steps are also included: Obtaining first location information of a distribution network device and second location information of multiple candidate drones; Determine the distance information between each of the candidate drones and the network distribution device according to the first location information of the network distribution device and the second location information of multiple candidate drones; From the candidate drones, a candidate drone with the smallest distance information to the network distribution device is selected as the drone to be controlled.

7. A drone control device, characterized in that: The device comprises: A video acquisition module is used to acquire the video to be analyzed that is taken by the drone to be controlled and is associated with the distribution network device; A video processing module, used for performing frame processing on the video to be analyzed to obtain multiple video frames to be analyzed; The video frame conversion module is used to convert each video frame to be analyzed to obtain text information corresponding to each video frame to be analyzed; A word vector acquisition module, used to obtain the word vector of the text information corresponding to each video frame to be analyzed; A word vector fusion module is used to fuse the word vectors of the text information corresponding to each video frame to be analyzed to obtain a fused word vector; An instruction prediction module, used to input the fused word vector into the trained UAV control instruction prediction model to obtain the target UAV control instruction corresponding to the fused word vector; The drone control module is used to control the drone to be controlled to monitor the distribution network equipment according to the target drone control instruction.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images