Unmanned aerial vehicle data transmission control method and system

Abstract

The invention discloses an unmanned aerial vehicle data transmission control method and system, which is applied to the technical field of line patrol distribution management, and comprises the following steps: obtaining unmanned aerial vehicle return image data in all patrol lines; performing duplicate judgment on all the returned image data to obtain each piece of to-be-transmitted inspection image data; pixel information in each piece of to-be-transmitted inspection image data is extracted, and a gradient energy index corresponding to each piece of to-be-transmitted inspection image data is obtained based on the pixel information; fusing the gradient energy indexes and the corresponding direction information to obtain a high-dimensional representation result of all the to-be-transmitted inspection image data; determining an estimated transmission state of each inspection image data to be transmitted according to each high-dimensional representation result; and when it is detected that the estimated transmission state meets the control condition, entering a corresponding control mode to execute a matched control action. According to the method, stable and efficient data return in the power grid equipment inspection process by the multiple unmanned aerial vehicles is ensured.

Description

Technical Field

[0001] This invention relates to the field of pipeline dispatch management technology, and in particular to a control method and system for UAV data transmission. Background Technology

[0002] As the core carrier for the stable operation of the power system, regular and precise inspection of power grid equipment is a necessary prerequisite for ensuring the safe and reliable operation of the power grid.

[0003] Current technologies utilize ground data backhaul interfaces to directly receive all image data collected by drones. However, this method does not consider issues such as link bandwidth load, data loss, or data delay during data backhaul, which directly impacts the continuity and reliability of power grid inspections. With increasing demands for safe power grid operation and the deployment of drones in large numbers and tiers, efficiently receiving image data has become a pressing issue for stable power grid operation. Summary of the Invention

[0004] This invention provides a control method and system for UAV data transmission to solve the technical problem that existing technologies cannot meet the requirements of large-scale, high-timeliness inspections, so as to achieve the high reliability and high efficiency requirements of power grid equipment inspection.

[0005] To address the aforementioned technical problems, the present invention provides a control method for UAV data transmission, comprising: Acquire image data transmitted back from drones along all inspection routes; Based on the direction information of each inspection route and the image content information extracted from each of the returned image data, all the returned image data are deduplicated to obtain each inspection image data to be transmitted. Extract pixel information from each of the inspection image data to be transmitted, and obtain a gradient energy index corresponding to each of the inspection image data to be transmitted based on the pixel information. The gradient energy index and the corresponding direction information are fused to obtain a high-dimensional representation of all the inspection image data to be transmitted; Based on the various high-dimensional representation results, the estimated transmission status of each of the inspection image data to be transmitted is determined; When the estimated transmission status is detected to meet the control conditions, the corresponding control mode is entered to execute the matching control action. The control action includes at least: transmitting each of the inspection image data to be transmitted sequentially to the ground data backhaul interface or sending the regenerated UAV inspection control signal to the selected UAV.

[0006] As one preferred embodiment, determining the estimated transmission status of each of the inspection image data to be transmitted based on the respective high-dimensional representation results includes: A preset transmission threshold is determined based on the initial transmission bandwidth of each of the aforementioned drones; Extract the texture index quantization component and density index quantization component from each of the high-dimensional representation results; The texture index quantization component and the density index quantization component are processed to obtain the total quantized value of each of the inspection image data to be transmitted. If the total quantified value is less than the preset transmission threshold, the estimated transmission status of the corresponding inspection image data to be transmitted is normal; otherwise, the estimated transmission status is abnormal.

[0007] As one preferred embodiment, the control mode includes at least a priority mode or a parallax adjustment mode; When the estimated transmission state is detected to meet the control conditions, the system enters the corresponding control mode to execute matching control actions, including: When a conflict is detected in the estimated transmission state, the available bandwidth margin of each of the drones is obtained. If the available bandwidth margin meets the preset conditions, the priority mode is entered. In the priority mode, the priority weight of each UAV is determined based on the available bandwidth margin, and the parameters of each inspection image data to be transmitted are optimized based at least on the gradient energy index of each inspection image data to be transmitted. Based on the priority weight, the optimized inspection image data to be transmitted are sequentially transmitted to the ground data backhaul interface; or; When a fault or crash is detected in the estimated transmission state, control information corresponding to the UAV is obtained. If the control information meets preset conditions, the parallax adjustment mode is entered. In the parallax adjustment mode, the actual shooting angle information in the control information is extracted, and the UAV inspection control signal is adjusted based at least on the actual shooting angle information. The regenerated UAV inspection control signal is then sent to the selected UAV.

[0008] As one preferred embodiment, in the priority mode, determining the priority weight of each UAV based on the available bandwidth margin includes: The quantitative indicators for determining the available bandwidth margin include at least: available bandwidth margin, transmission queue length, and estimated transmission duration; Each of the aforementioned quantitative indicators is analyzed to obtain a quantitative score; The quantified scores are weighted and summed to obtain the priority weight for each UAV.

[0009] As one preferred embodiment, the step of determining duplicates in all the returned image data based on the direction information of each of the inspection routes and the image content information extracted from each of the returned image data to obtain each inspection image data to be transmitted includes: All the returned image data are subjected to first-level deduplication, specifically including: Acquire all historical image data and location information corresponding to all historical image data on each inspection route; extract historical image content information of all historical image data. The historical image content information is subjected to the first-level deduplication, and the location information corresponding to the duplicate historical image data is marked; The returned image data corresponding to the location information is removed to obtain the first-level deduplication result.

[0010] As one preferred embodiment, the step of removing the corresponding location information from the returned image data to obtain the first-level deduplication result further includes: A two-level deduplication process is performed on all the returned image data, specifically including: Extract all image content information from the first-level deduplication result to obtain a feature set; The feature set is subjected to the second-level deduplication to obtain each of the inspection image data to be transmitted. The second-level deduplication includes analyzing the content information of each of the images in the feature set to obtain the similarity of the corresponding image content information.

[0011] As one preferred embodiment, the extraction of pixel information from each of the inspection image data to be transmitted includes: Each of the inspection image data to be transmitted is processed into grayscale to obtain grayscale image data of each of the inspection image data to be transmitted. Extract the grayscale value matrix of each grayscale image data; The grayscale value matrix of each grayscale image data is processed to obtain the pixel information of each inspection image data to be transmitted.

[0012] As one preferred embodiment, obtaining the gradient energy index corresponding to each of the inspected image data to be transmitted based on the pixel information includes: Obtain the grayscale value matrix from all the pixel information; Extract the horizontal and vertical gradient magnitudes of each of the grayscale value matrices; Based on the gradient magnitudes of each horizontal direction and each vertical direction, the gradient energy value of each of the inspection image data to be transmitted is obtained. The gradient energy value is normalized to obtain the gradient energy index for each of the inspection image data to be transmitted.

[0013] As a preferred embodiment, the step of fusing the gradient energy index and the corresponding direction information to obtain a high-dimensional representation of all the inspection image data to be transmitted includes: The gradient energy index of each of the inspection image data to be transmitted is processed to obtain the gradient energy index product. Spatial mapping is performed on the gradient energy index product and the direction information to obtain a gradient energy index vector and a direction vector. The gradient energy index vector and the direction vector are then fused to obtain a high-dimensional feature vector. The high-dimensional feature vector is standardized to obtain the high-dimensional representation of all the inspection image data to be transmitted.

[0014] Another aspect of the present invention provides a control system for unmanned aerial vehicle (UAV) data transmission, comprising: The data acquisition module is used to acquire the image data transmitted back by the drones in all inspection routes; The data deduplication module is used to deduplicatively evaluate all the returned image data based on the direction information of each of the inspection routes and the image content information extracted from each of the returned image data, so as to obtain each inspection image data to be transmitted. The data preprocessing module is used to extract pixel information from each of the inspection images to be transmitted, and to obtain a gradient energy index corresponding to each of the inspection images to be transmitted based on the pixel information. The data fusion module is used to fuse the gradient energy index and the corresponding direction information to obtain a high-dimensional representation of all the inspection image data to be transmitted. The status analysis module is used to determine the estimated transmission status of each of the inspection image data to be transmitted based on the results of each of the high-dimensional representations. The status control module is used to enter the corresponding control mode to execute matching control actions when the estimated transmission status is detected to meet the control conditions. The control actions include at least: sequentially transmitting each of the inspection image data to be transmitted to the ground data backhaul interface or sending the regenerated UAV inspection control signal to the selected UAV.

[0015] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: The present invention uses the image gradient energy index and corresponding direction information of multi-UAV collaborative inspection to obtain the high-dimensional representation result of the inspection image data to be transmitted, and uses the high-dimensional representation result as the judgment index of data transmission status to accurately identify the transmission status on the transmission link. Combined with the control mode, it realizes flexible and efficient multi-UAV image data transmission, avoids the problem of blind data transmission leading to transmission link overload, and ensures stable and efficient data backhaul during the inspection of power grid equipment by multiple UAVs. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a control method for UAV data transmission in one embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a UAV data transmission control system in one embodiment of the present invention; Figure label: The modules are: 11. Data acquisition module; 12. Data deduplication module; 13. Data preprocessing module; 14. Data fusion module; 15. Status analysis module; and 16. Status control module. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] The terms “vertical,” “horizontal,” “left,” “right,” “up,” “down,” and similar expressions used herein are for illustrative purposes only and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention. The term “and / or” as used herein includes any and all combinations of one or more of the associated listed items. Those skilled in the art will understand the specific meaning of the above terms in this application based on the specific circumstances.

[0019] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0020] In existing technologies, directly receiving image data collected by drones via a ground data backhaul interface is a common method. Its operational logic is relatively straightforward: after the drone completes image data collection, it wirelessly transmits this data directly to the ground data backhaul interface. However, during data backhaul, when the amount of image data collected by the drone exceeds the bandwidth capacity of the transmission link, the data transmission speed is limited, leading to data transmission interruptions, data loss, or delayed reception. With the continuous expansion of the power grid and its increasing complexity, a large number of drones are being deployed in batches and tiers to meet the growing demand for power grid inspection. However, existing technologies that directly receive all image data collected by drones via a ground data backhaul interface are inadequate for such large-scale data transmission. Ensuring efficient image data reception has become an urgent problem to be solved for the stable operation of the power grid.

[0021] Furthermore, in this embodiment of the invention, in order to achieve the goal of comprehensive and thorough inspection of power grid equipment, multiple drones with complementary functions are allocated to each inspection route, forming a tiered inspection unit. The inspection routes include high-voltage transmission lines, substation perimeter lines, and grounding device lines, etc., and the drones include drones equipped with high-definition optical cameras, drones equipped with infrared thermal imagers, and drones equipped with lidar, etc. Simultaneously, to maximize the utilization rate of drone resources and adapt to the "numerous points, long lines, and wide coverage" distribution characteristics of power grid inspection routes, each drone undertakes the operation tasks of multiple adjacent inspection routes through a dynamic task scheduling mechanism.

[0022] For details, please see Figure 1 , Figure 1 The diagram shown is a flowchart illustrating a control method for UAV data transmission in one embodiment of the present invention. The method includes steps S1 to S6: S1. Obtain the image data transmitted back by the drones in all inspection routes; S2. Based on the direction information of each inspection route and the image content information extracted from each of the returned image data, perform duplicate judgment on all the returned image data to obtain each inspection image data to be transmitted. S3. Extract pixel information from each of the inspection image data to be transmitted, and obtain the gradient energy index corresponding to each of the inspection image data to be transmitted based on the pixel information. S4. The gradient energy index and the corresponding direction information are fused to obtain a high-dimensional representation of all the inspection image data to be transmitted; S5. Based on the high-dimensional representation results, determine the estimated transmission status of each of the inspection image data to be transmitted. S6. When the estimated transmission status is detected to meet the control conditions, the corresponding control mode is entered to execute the matching control action. The control action includes at least: transmitting each of the inspection image data to be transmitted sequentially to the ground data backhaul interface or sending the regenerated UAV inspection control signal to the selected UAV.

[0023] Furthermore, in step S1, before the UAVs perform the inspection task, the ground dispatch system performs communication pairing for all UAVs, supporting parameter distribution and synchronization with the inspection task. The UAVs determine the optimal inspection trajectory based on the basic information of the inspection route and historical inspection data. Image data acquisition trigger modes are set, including automatic and manual acquisition triggers. In the automatic acquisition trigger module, the automatic shooting time is set. In the manual acquisition trigger module, staff can use the remote control interface of the ground dispatch system to configure a manual reshoot button for each UAV.

[0024] Furthermore, in step S2, during the UAV's inspection mission, a large amount of image data collected is transmitted back. This image data plays a crucial role in accurately assessing the status of the inspected object and promptly identifying potential problems. However, due to various factors during the inspection process, such as shooting from different angles and monitoring repeated areas, the transmitted image data often contains a large amount of duplicate image data. This duplicate image data not only occupies storage space but also increases the complexity of data management and consumes a significant amount of network bandwidth and time during data transmission, affecting data transmission efficiency. Therefore, it is essential to perform deduplication processing on all transmitted image data and extract representative inspection image data to be transmitted.

[0025] Specifically, a first-level deduplication process is required for all transmitted image data. This includes retrieving historical image data for each inspection route and the corresponding location information from the historical database of the UAV inspection system. A feature recognition algorithm is then used to extract the historical image content information from all historical image data. This historical image content information is then subjected to a first-level deduplication process, marking the location information corresponding to duplicate historical image data and removing the corresponding location information from the transmitted image data. This yields the first-level deduplication result, thus achieving redundancy removal within the historical image data. Secondly, a second-level deduplication process is performed on the transmitted image data based on the first-level deduplication result. This includes extracting all image content information from the first-level deduplication result to obtain a feature set. The feature set is then subjected to a second-level deduplication process to obtain the inspection image data to be transmitted. Preferably, the second-level deduplication process involves analyzing the image content information in the feature set to obtain the similarity of the corresponding image content information. If the similarity exceeds a preset threshold, the image data is determined to be duplicated and is removed.

[0026] Furthermore, in step S3, pixel information determines the resolution and sharpness of the image. The more pixel data, the higher the image resolution, the richer the details, and the more accurately the image can present the original scene. The gradient energy index is used to measure the degree of pixel change and the richness of detail in the image. A high gradient energy index means that the image contains rich details and sharp edges, and the image quality is high. Conversely, a low gradient energy index means that the image may be blurry and has lost a lot of detail.

[0027] Specifically, firstly, each image data to be transmitted for inspection is processed into grayscale to generate corresponding grayscale image data. Then, the grayscale value matrix of each grayscale image data is extracted—this matrix is ​​the core digital representation of the grayscale image. In image feature analysis, gradient magnitude accurately reflects the rate of change of pixel grayscale values ​​in an image and is a key indicator for measuring the richness of image edges and details. Therefore, for each grayscale value matrix, the corresponding horizontal and vertical gradient magnitudes need to be extracted. The gradient energy value, as a global statistical result of the gradient magnitude, can quantitatively characterize the richness of edges and details in the entire image. Specifically, the horizontal and vertical gradient magnitudes of each grayscale value matrix are summed to obtain the gradient energy value for each image data to be transmitted for inspection. Finally, the gradient energy value is normalized to obtain the gradient energy index for each image data to be transmitted for inspection.

[0028] Furthermore, in step S4, the high-dimensional representation result can extract key features from complex image data and map the data to a high-dimensional space for in-depth analysis, thereby keenly capturing subtle changes in the data transmission process and laying a solid foundation for subsequent data transmission status determination. Specifically, firstly, the gradient energy index of each image data to be transmitted is nonlinearly transformed to obtain the gradient energy index product. Then, the radial basis function (RBF) is used to spatially map the gradient energy index product with the direction information to generate a gradient energy index vector and a direction vector. Next, these two types of vectors are concatenated and fused to obtain a high-dimensional feature vector. Finally, the Min-Max normalization method is used to standardize the high-dimensional feature vector to obtain the high-dimensional representation result of all the image data to be transmitted.

[0029] Furthermore, in step S5, determining the estimated transmission status is a crucial step in ensuring the smooth completion of the entire inspection task and subsequent data analysis and application. From the perspective of ensuring data transmission integrity, if an abnormal transmission status occurs, leading to data loss or corruption, the analysis and decisions made based on this incomplete data may be seriously flawed. For example, in power line inspections, if image data of critical areas is missing due to transmission anomalies, it becomes impossible to accurately determine whether the line is damaged or aging, thus posing a potential risk to the stable operation of the power system. By determining the estimated transmission status, problems that may affect data integrity can be identified in advance, allowing for timely correction or retransmission, ensuring that the ground data interface receives complete and accurate image data, providing a reliable basis for subsequent data analysis and fault diagnosis.

[0030] Specifically, the texture index quantization component and density index quantization component are extracted from each high-dimensional representation result. These components are then weighted based on preset weights to obtain a total quantized value. This total quantized value is compared with a preset transmission threshold. If the total quantized value is less than the threshold, the estimated transmission status of the corresponding inspection image data is normal; otherwise, the estimated transmission status is abnormal. The texture index quantization component refers to the extraction of texture features from the high-dimensional representation result through local binary mode transformation, and the calculation of the root mean square (RMS) of all local binary mode values, which is then used as the texture index. The density index quantization component represents the spatial clustering degree of non-zero elements in the high-dimensional representation result. The initial transmission bandwidth of the UAV is crucial for determining the preset transmission threshold. The method for obtaining the initial transmission bandwidth typically depends on the communication module carried by the UAV and the characteristics of its communication link with the ground data interface. Preferably, the preset weights can be set based on human experience.

[0031] Furthermore, in step S6, if the estimated transmission state is detected to meet the control conditions, the corresponding control mode is entered to execute the matching control action to ensure the stability, efficiency, and security of the data transmission process. Among them, the setting of control conditions needs to cover the key abnormal scenarios that may occur during the transmission process, including at least the following two core situations: (1) When there is a conflict in the estimated transmission state, that is, multiple UAVs simultaneously initiate data transmission requests to the ground data backhaul interface, resulting in problems such as transmission resource competition, data transmission queue congestion, or transmission link occupation conflict; (2) When there is a failure or collapse in the estimated transmission state, that is, the transmission link between a single or multiple UAVs and the ground data backhaul interface is interrupted, the signal is severely attenuated, the data packet loss rate exceeds the threshold, or the data transmission module of the UAV itself is abnormal, resulting in a collapse scenario in which data transmission cannot proceed normally.

[0032] If a conflict is detected in the estimated transmission status, the available bandwidth margin of each UAV is first obtained (i.e., the amount of bandwidth resources that are not currently occupied by the UAV and can be used for the transmission of inspection image data). The obtained available bandwidth margin of each UAV is compared with the preset bandwidth threshold condition (this threshold is preset based on factors such as the maximum carrying bandwidth of the ground data backhaul interface and the minimum transmission bandwidth requirement of the inspection image data). If the available bandwidth margin meets the preset condition, the priority mode is entered. Specifically, in the priority mode, the priority weight of each UAV is determined based on the available bandwidth margin. The larger the available bandwidth margin, the higher the priority of the corresponding UAV. At least based on the gradient energy index of each inspection image data to be transmitted, the parameters of each inspection image data to be transmitted are optimized. Based on the priority weight, the optimized inspection image data to be transmitted is transmitted to the ground data backhaul interface in sequence. Preferably, the parameter optimization process includes calculating the scaling factor (i.e., the compression ratio of the image size; images with low gradient energy can use a larger scaling factor to reduce the amount of data, while images with high gradient energy can use a smaller scaling factor to preserve details) and the feature quantization index (i.e., the precision parameter when quantizing image feature data; images with rich details use higher quantization precision, while images with simple details can appropriately reduce quantization precision) and the feature quantization index through coordinated adjustment of the scaling factor and the feature quantization index, so as to minimize the amount of data transmitted while ensuring that key image information is not lost.

[0033] If a failure or crash is detected in the estimated transmission status, the corresponding control information for the UAV is first obtained through the UAV status monitoring system. If the control information meets preset conditions, the parallax adjustment mode is entered. This control information includes at least the UAV's real-time position information, flight attitude data, data transmission module operating status, camera parameters, and historical fault data. Specifically, in parallax adjustment mode, the actual shooting angle information of the control information is extracted, and the UAV inspection control signal is adjusted based on at least this actual shooting angle information. The regenerated UAV inspection control signal is then sent to the selected UAV.

[0034] Another embodiment of the present invention provides a control system for UAV data transmission; for details, please refer to [link to relevant documentation]. Figure 2 , Figure 2 The diagram shown illustrates the structure of a UAV data transmission control system according to one embodiment of the present invention. The system includes: Data acquisition module 11 is used to acquire the image data transmitted back by drones in all inspection routes; The data deduplication module 12 is used to deduplicatively evaluate all the returned image data based on the direction information of each of the inspection lines and the image content information extracted from each of the returned image data, so as to obtain each inspection image data to be transmitted. Data preprocessing module 13 is used to extract pixel information from each of the inspection image data to be transmitted, and to obtain a gradient energy index corresponding to each of the inspection image data to be transmitted based on the pixel information. Data fusion module 14 is used to fuse the gradient energy index and the corresponding direction information to obtain a high-dimensional representation of all the inspection image data to be transmitted; The state analysis module 15 is used to determine the estimated transmission state of each of the inspection image data to be transmitted based on the high-dimensional representation results. The status control module 16 is used to enter the corresponding control mode to execute matching control actions when the estimated transmission status is detected to meet the control conditions. The control actions include at least: transmitting each of the inspection image data to be transmitted sequentially to the ground data backhaul interface or sending the regenerated UAV inspection control signal to the selected UAV.

[0035] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: (1) This invention analyzes the image gradient energy index and corresponding direction information of multi-UAV collaborative inspection, constructs a high-dimensional representation of the inspection image data to be transmitted, and directly uses the high-dimensional representation as the judgment index of data transmission status, so as to comprehensively and accurately reflect the transmission status on the transmission link. (2) The present invention further combines control modes to realize dynamic optimization of multi-UAV image data transmission. The corresponding control mode is triggered for different transmission states to realize flexible and efficient multi-UAV image data transmission, avoiding the problem of overload of transmission links caused by blind data transmission, and ensuring stable and efficient data back transmission of multiple UAVs during the inspection of power grid equipment.

[0036] (3) While avoiding the risk of transmission link overload, this invention improves data transmission efficiency through priority scheduling, data parameter optimization and other means, and ensures that the massive image data generated during the inspection process can be promptly transmitted back to the ground data transmission interface, providing timely data support for power grid equipment fault analysis and status assessment.

[0037] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A control method for data transmission in an unmanned aerial vehicle (UAV), characterized in that, include: Acquire image data transmitted back from drones along all inspection routes; Based on the direction information of each inspection route and the image content information extracted from each of the returned image data, all the returned image data are deduplicated to obtain each inspection image data to be transmitted. Extract pixel information from each of the inspection image data to be transmitted, and obtain a gradient energy index corresponding to each of the inspection image data to be transmitted based on the pixel information. The gradient energy index and the corresponding direction information are fused to obtain a high-dimensional representation of all the inspection image data to be transmitted. Based on the various high-dimensional representation results, the estimated transmission status of each of the inspection image data to be transmitted is determined; When the estimated transmission status is detected to meet the control conditions, the corresponding control mode is entered to execute the matching control action. The control action includes at least: transmitting each of the inspection image data to be transmitted sequentially to the ground data backhaul interface or sending the regenerated UAV inspection control signal to the selected UAV.

2. The control method for UAV data transmission as described in claim 1, characterized in that, The step of determining the estimated transmission status of each of the inspection image data to be transmitted based on the various high-dimensional representation results includes: A preset transmission threshold is determined based on the initial transmission bandwidth of each of the aforementioned drones; Extract the texture index quantization component and density index quantization component from each of the high-dimensional representation results; The texture index quantization component and the density index quantization component are processed to obtain the total quantized value of each of the inspection image data to be transmitted. If the total quantified value is less than the preset transmission threshold, the estimated transmission status of the corresponding inspection image data to be transmitted is normal; otherwise, the estimated transmission status is abnormal.

3. The control method for UAV data transmission as described in claim 1, characterized in that, The control mode includes at least a priority mode or a parallax adjustment mode; When the estimated transmission state is detected to meet the control conditions, the system enters the corresponding control mode to execute matching control actions, including: When a conflict is detected in the estimated transmission state, the available bandwidth margin of each of the drones is obtained. If the available bandwidth margin meets the preset conditions, the priority mode is entered. In the priority mode, the priority weight of each UAV is determined based on the available bandwidth margin, and the parameters of each inspection image data to be transmitted are optimized based at least on the gradient energy index of each inspection image data to be transmitted. Based on the priority weight, the optimized inspection image data to be transmitted are sequentially transmitted to the ground data backhaul interface; or; When a fault or crash is detected in the estimated transmission state, control information corresponding to the UAV is obtained. If the control information meets preset conditions, the parallax adjustment mode is entered. In the parallax adjustment mode, the actual shooting angle information in the control information is extracted, and the UAV inspection control signal is adjusted based at least on the actual shooting angle information. The regenerated UAV inspection control signal is then sent to the selected UAV.

4. The control method for UAV data transmission as described in claim 3, characterized in that, In the priority mode, determining the priority weight of each UAV based on the available bandwidth margin includes: The quantitative indicators for determining the available bandwidth margin include at least: available bandwidth margin, transmission queue length, and estimated transmission duration; Each of the aforementioned quantitative indicators is analyzed to obtain a quantitative score; The quantified scores are weighted and summed to obtain the priority weight for each UAV.

5. The control method for UAV data transmission as described in claim 1, characterized in that, The step of determining duplicates in all the returned image data based on the direction information of each of the inspection routes and the image content information extracted from each of the returned image data to obtain each inspection image data to be transmitted includes: All the returned image data are subjected to first-level deduplication, specifically including: Acquire all historical image data and location information corresponding to all historical image data on each inspection route; extract historical image content information of all historical image data. The historical image content information is subjected to the first-level deduplication, and the location information corresponding to the duplicate historical image data is marked; The returned image data corresponding to the location information is removed to obtain the first-level deduplication result.

6. The control method for UAV data transmission as described in claim 5, characterized in that, The step of removing the corresponding location information from the returned image data to obtain the first-level deduplication result also includes: A two-level deduplication process is performed on all the returned image data, specifically including: Extract all image content information from the first-level deduplication result to obtain a feature set; The feature set is subjected to the second-level deduplication to obtain each of the inspection image data to be transmitted. The second-level deduplication includes analyzing the content information of each of the images in the feature set to obtain the similarity of the corresponding image content information.

7. The control method for UAV data transmission as described in claim 1, characterized in that, The step of extracting pixel information from each of the inspection image data to be transmitted includes: Each of the inspection image data to be transmitted is processed into grayscale to obtain grayscale image data of each of the inspection image data to be transmitted. Extract the grayscale value matrix of each grayscale image data; The grayscale value matrix of each grayscale image data is processed to obtain the pixel information of each image data to be transmitted for inspection.

8. The control method for UAV data transmission as described in claim 7, characterized in that, The step of obtaining the gradient energy index corresponding to each of the inspection image data to be transmitted based on the pixel information includes: Obtain the grayscale value matrix from all the pixel information; Extract the horizontal and vertical gradient magnitudes of each of the grayscale value matrices; Based on the gradient magnitudes of each horizontal direction and each vertical direction, the gradient energy value of each of the inspection image data to be transmitted is obtained. The gradient energy value is normalized to obtain the gradient energy index for each of the inspection image data to be transmitted.

9. The control method for UAV data transmission as described in claim 1, characterized in that, The step of fusing the gradient energy index and the corresponding direction information to obtain a high-dimensional representation of all the inspection image data to be transmitted includes: The gradient energy index of each of the inspection image data to be transmitted is processed to obtain the gradient energy index product. Spatial mapping is performed on the gradient energy index product and the direction information to obtain a gradient energy index vector and a direction vector. The gradient energy index vector and the direction vector are then fused to obtain a high-dimensional feature vector. The high-dimensional feature vector is standardized to obtain the high-dimensional representation of all the inspection image data to be transmitted.

10. A control system for data transmission from an unmanned aerial vehicle (UAV), characterized in that, include: The data acquisition module is used to acquire the image data transmitted back by the drones in all inspection routes; The data deduplication module is used to deduplicatively evaluate all the returned image data based on the direction information of each of the inspection routes and the image content information extracted from each of the returned image data, so as to obtain each inspection image data to be transmitted. The data preprocessing module is used to extract pixel information from each of the inspection images to be transmitted, and to obtain a gradient energy index corresponding to each of the inspection images to be transmitted based on the pixel information. The data fusion module is used to fuse the gradient energy index and the corresponding direction information to obtain a high-dimensional representation of all the inspection image data to be transmitted. The status analysis module is used to determine the estimated transmission status of each of the inspection image data to be transmitted based on the results of each of the high-dimensional representations. The status control module is used to enter the corresponding control mode to execute matching control actions when the estimated transmission status is detected to meet the control conditions. The control actions include at least: sequentially transmitting each of the inspection image data to be transmitted to the ground data backhaul interface or sending the regenerated UAV inspection control signal to the selected UAV.

Images