Automatic planning system for shooting waypoints of electric unmanned aerial vehicle

Abstract

The invention discloses an electric unmanned aerial vehicle shooting waypoint automatic planning system, and relates to the technical field of path planning, the system comprises the following components: a data acquisition and integration module, a data analysis and intelligent identification module, a waypoint intelligent planning module, a dynamic adjustment decision module and a communication and data transmission module; according to the method, the electric power facility data before and after disasters are integrated, deep analysis is carried out by using an image recognition technology and a machine learning algorithm, the time for acquiring the image data of the damage condition of the electric power facilities is greatly shortened, the comprehensiveness and accuracy of the data are ensured, and the system preferentially shoots key equipment, so that the system is convenient to use. A low-altitude multi-angle shooting strategy is adopted, and detailed and reliable image data is provided for power line repair work, so that the overall efficiency of emergency assessment and the scientificity of repair decision are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and in particular to an automatic waypoint planning system for shooting electric unmanned aerial vehicles. Background Art

[0002] After various natural disasters such as earthquakes, floods, and typhoons, power facilities often suffer varying degrees of damage. Rapidly and accurately assessing the damage to power facilities is crucial for subsequent power repair work, as it is directly related to the speed and efficiency of power restoration. With the rapid development of drone technology, using drones to photograph and assess affected power facilities has become an efficient and relatively safe means. Drones, with their strong maneuverability, wide coverage, and flexible shooting angles, are playing an increasingly important role in disaster assessment.

[0003] Although drones have shown great potential in emergency assessments of power facilities, there are still many shortcomings in the current drone photography waypoint planning. On the one hand, traditional methods lack scientific, systematic and intelligent planning means, and cannot fully utilize pre-disaster power facility data and rich post-disaster information, resulting in unreasonable waypoint planning and low photography efficiency. On the other hand, traditional technology is difficult to automatically and accurately avoid dangerous areas, such as leakage areas, unstable building areas, etc., and it is also difficult to flexibly adjust waypoint planning according to the actual progress of rescue and the specific needs of power repair. These problems make the acquired image data of power facility damage neither comprehensive nor accurate, seriously hindering the restoration process of power facilities, and may even cause assessors to face unnecessary risks.

[0004] In response to the above problems, it is necessary to optimize the existing automatic planning system for electric power drone photography waypoints. By comprehensively collecting and integrating pre-disaster power facility data and post-disaster information, using image recognition technology and machine learning algorithms for in-depth analysis, and combining the drone's flight performance parameters and dangerous area information, a scientific and reasonable emergency assessment waypoint planning can be generated. Therefore, it is of great significance to develop an automatic planning system for electric power drone photography waypoints that can comprehensively realize the above characteristics. Summary of the Invention

[0005] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide an automatic planning system for waypoints for shooting electric drones. It can comprehensively collect pre-disaster power facility data and post-disaster remote sensing images, on-site videos and other information through the data acquisition and integration module, and use data mapping algorithms for precise mapping and deep fusion. Through the data analysis and intelligent recognition module, image recognition technology and machine learning algorithms are used to conduct in-depth analysis of the integrated data, accurately identify the collapse of buildings, the possible damaged locations of power equipment, and the specific location, range and danger level of dangerous areas. Based on these precise results, the waypoint intelligent planning module uses an intelligent path planning algorithm to quickly generate a scientific and reasonable emergency assessment waypoint plan, and fully considers the flight performance parameters of the drone and the dangerous area information to automatically and accurately avoid the dangerous area. At the same time, the dynamic adjustment decision module can receive dynamic information on rescue progress and power repair needs in real time, and make dynamic adjustments to the waypoint planning in a timely manner to ensure that drone shooting can always closely meet the actual needs of power repair.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an automatic waypoint planning system for shooting electric UAVs, the system comprising the following components: a data acquisition and integration module, a data analysis and intelligent recognition module, a waypoint intelligent planning module, a dynamic adjustment decision module, and a communication and data transmission module;

[0007] The data collection and integration module is responsible for collecting and integrating pre-disaster and post-disaster data, and establishing a mapping relationship between pre-disaster and post-disaster data. It minimizes the error through iterative optimization, fuses the mapped data, builds a unified structure and stores it in the database management system.

[0008] The data analysis and intelligent identification module uses image segmentation technology to pre-process post-disaster remote sensing images and videos, identify the locations of collapsed buildings and damaged power equipment, collect a large amount of dangerous area sample data, annotate and classify them, build a machine learning classification model to identify and classify dangerous areas, and optimize the model through cross-validation and evaluation indicators;

[0009] The intelligent waypoint planning module determines a list of key equipment based on the power system topology and operating rules and assigns high priority to it. It then prioritizes the waypoints it uses to shoot its equipment. It also determines the low-altitude flight and multi-angle shooting range based on the drone's performance and shooting requirements, designs a path to obtain a comprehensive image, and generates initial waypoints based on hazardous area information and shooting strategies.

[0010] The dynamic adjustment decision module sets rescue progress and emergency repair demand indicators, triggering the adjustment mechanism when the indicators change. By establishing a real-time communication interface with the emergency repair command center and on-site personnel, it obtains and analyzes data, determines the adjustment range of the waypoint, and dynamically adjusts the initial waypoint plan;

[0011] The communication and data transmission module performs two-way communication through a wireless communication protocol, formulates a unified transmission format to process waypoint planning and image data, encrypts the data, establishes a communication link monitoring mechanism to detect transmission reliability in real time, and automatically switches or reconnects in case of anomalies.

[0012] Furthermore, the data collection and integration module is responsible for the collection and integration of pre-disaster and post-disaster data. Before the disaster, the location information of power facilities is obtained from the power enterprise resource management system and GIS, and the structural diagram is extracted from the design database. After the disaster, remote sensing images are collected using a satellite remote sensing platform, and on-site personnel are organized to collect videos using mobile terminals. A mapping relationship between pre-disaster and post-disaster data is established. The matching degree formula of the mapping relationship is: ,in, Indicates the mapping matching degree between pre-disaster and post-disaster data. The smaller the value, the higher the matching degree. is the number of data features, No. The weight of the feature, This is the pre-disaster data eigenvalues ​​from the power company's historical database, design drawings, and geographic information system records. This is the first post-disaster data eigenvalues, obtained through post-disaster remote sensing image processing, on-site video analysis, and field survey data extraction. is a supplementary term to avoid the denominator being zero.

[0013] Furthermore, the data analysis and intelligent recognition module uses image segmentation technology to pre-process post-disaster remote sensing images and videos. Specifically, it imports image and video data and converts and standardizes their formats. It uses the histogram equalization method to enhance image contrast and filter out noise, and extracts spectral, texture and geometric features. It uses threshold-based, region-based and machine learning segmentation methods to divide the image into different regions. Through morphological operations and region merging and screening processing, the segmentation results are optimized.

[0014] Furthermore, the data analysis and intelligent identification module constructs a machine learning classification model to identify and classify dangerous areas. The model formula is: ,in, Is the loss function, which is used to measure the difference between the model prediction result and the actual label. is the number of training samples, It is The true label of the sample, 0 represents a non-dangerous area, 1 represents a dangerous area, The model is The probability that a sample is predicted to be a dangerous area, is the parameter vector of the model, is the number of model parameters, yes The regularization term sums the absolute values ​​of the model parameters so that some parameter values ​​become zero, thereby achieving the purpose of feature selection. yes The regularization term, also known as weight decay, makes the parameter value tend to be smaller by summing the squares of the model parameters, thereby reducing the complexity of the model. and Regularization terms and The coefficients are optimized and determined during the training process through the cross-validation method.

[0015] Furthermore, the waypoint intelligent planning module determines a list of key equipment based on the power system topology and operating rules, assigns high priority, and prioritizes waypoint planning for these equipment. Specifically, the topological connection relationship data and operating rule data of the power system are obtained. Based on this data, the power system is abstracted into a topological graph. By analyzing the connectivity of the graph, nodes and connection paths where faults may cause large-scale power outages or affect important power supply areas are identified. At the same time, the power distribution during normal operation of the power system is analyzed to determine the equipment and lines that carry large power transmission tasks. Based on the rated parameters of the equipment, large-capacity generators and main transformers are screened. Combined with the load rate and the protection requirements of important users, key equipment for long-term high-load operation and power supply to important users is determined to form a list of key equipment and assign high-priority identification to the key equipment in the list. The shooting area is determined based on the equipment location coordinates and the drone shooting angle and resolution requirements. An intelligent path planning algorithm is used to generate waypoints within the shooting area that meet the "low-altitude multi-angle shooting" strategy. When generating waypoints, the drone's flight performance limitations and dangerous area information are taken into account. Shooting waypoints are planned for high-priority key equipment in order of priority.

[0016] Furthermore, the waypoint intelligent planning module uses an intelligent path planning algorithm to generate waypoints that meet the "low-altitude multi-angle shooting" strategy within the shooting area. The algorithm formula is: ,in, is a node The estimated total cost of From the starting point to the node The actual flight cost is calculated based on the flight performance parameters and map information of the UAV. is a node The straight-line distance to the target waypoint is heuristically estimated using the waypoint's coordinate data. is a node The surrounding danger level assessment value is calculated based on the distribution and intensity of the danger zone. and are the weight coefficients of distance heuristic and danger level, respectively, which are dynamically adjusted according to the safety and efficiency requirements of the flight mission.

[0017] Furthermore, the dynamic adjustment decision module sets rescue progress and repair demand indicators. For the rescue progress indicator, the equipment repair status is obtained from the power repair command center, the number of repairs is counted by type and compared with the total number before the disaster, the overall repair progress ratio is calculated by weight, the power supply area is divided according to the power topology and GIS data, the number of restored power supply areas is counted with the help of the power dispatching system, the power supply restoration ratio is obtained, and positioning equipment is provided for the repair team. The advancement position is determined in combination with the task point, and the task completion ratio is calculated to comprehensively measure the overall advancement index. For the power repair demand indicator, new damaged equipment information is collected through real-time image recognition and on-site feedback, classified and counted by type and degree of damage, and special repair needs are collected and quantified through communication with all parties. Dynamic information of dangerous areas is obtained from the data analysis and intelligent identification module, and its scope and level changes are analyzed and quantified. The historical data of disaster repair is analyzed to set the initial threshold for each indicator, and the threshold is dynamically optimized based on the indicator data in the actual rescue and the adjustment effect of the waypoint planning. For the rescue progress indicator, the adjustment formula is: ,in, is the change in rescue progress, and are the rescue progress indicators of the current and previous moments, obtained from the real-time statistical data of the power repair command center. It is the trigger threshold for changes in rescue progress. It is pre-set or dynamically adjusted according to the scale and complexity of the rescue mission and the timeliness of information updates. For the power repair demand indicator, the adjustment formula is: ,in, is the change in demand for power repairs, and The power repair demand indicators at the current and previous moments are determined based on the feedback from on-site repair personnel and the newly discovered damage. It is the trigger threshold for changes in power emergency repair needs, which is pre-set or dynamically adjusted according to the scale and complexity of the rescue mission and the timeliness of information updates.

[0018] Furthermore, the dynamic adjustment decision module acquires and analyzes data by establishing a real-time communication interface with the emergency repair command center and on-site personnel to determine the adjustment range of the waypoint. The adjustment range is: ,in, is the adjustment range of waypoint planning, and They are the impact coefficients of changes in rescue progress and power repair demand on the waypoint adjustment range, which are obtained through analysis of historical rescue data and machine learning model training to adapt to different types of disasters and rescue scenarios.

[0019] Furthermore, the communication and data transmission module establishes a communication link monitoring mechanism to detect transmission reliability in real time. The reliability monitoring formula is: ,in, is the reliability score of data transmission, The number of data frames successfully transmitted, is the total number of data frames sent, which is obtained from the transmission log statistics of the communication module. It is the weight of the transmission success rate, which is determined according to the real-time requirements and integrity requirements of data transmission. is the actual data transmission time, which is obtained through the timestamp record of the communication module. is the expected data transmission time, calculated based on the amount of data and the theoretical transmission rate of the communication protocol, The weight of transmission time compliance is determined based on the real-time requirements and integrity requirements of data transmission.

[0020] Compared with the existing technology, this electric UAV shooting waypoint automatic planning system has the following beneficial effects:

[0021] 1. This invention integrates pre- and post-disaster power facility data and uses image recognition technology and machine learning algorithms for in-depth analysis. This not only significantly shortens the time to obtain image data on power facility damage, but also ensures the comprehensiveness and accuracy of the data. The system prioritizes photographing key equipment and adopts a low-altitude, multi-angle shooting strategy to provide detailed and reliable image data for power repair work, thereby significantly improving the overall efficiency of emergency assessments and the scientific nature of repair decisions.

[0022] 2. The present invention can receive dynamic information on rescue progress and power repair needs in real time through a dynamic adjustment decision module, and timely adjust the waypoint planning according to the preset dynamic adjustment trigger mechanism, ensuring that the drone shooting operation can closely meet the actual needs of power repair. At the same time, the system can also automatically and accurately avoid dangerous areas, such as leakage areas, unstable building areas, etc., to ensure the safety of drone shooting operations, which not only improves shooting efficiency, but also ensures the life safety of evaluators.

[0023] Other advantages, objects and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be learned from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0025] Figure 1 This is a structural diagram of an automatic waypoint planning system for electric drone photography.

[0026] Figure 2 This is a flowchart of an automatic waypoint planning system for an electric UAV. DETAILED DESCRIPTION

[0027] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0028] Example 1: A strong earthquake struck a city, causing buildings to shake and collapse, causing serious damage to local power facilities. Many transmission line towers tilted or collapsed, some equipment in the substation shifted and was damaged, and some underground cables were also damaged by ground deformation, resulting in power outages in parts of the city.

[0029] Quickly obtain detailed data on power facilities before the disaster from the power company's management system, including the topology of the entire city's power grid, the equipment list of each substation, the laying path of power lines, the specific location and model parameters of towers, etc. At the same time, collect relevant geographic information data, such as topography, building distribution, etc. for subsequent analysis, use satellite remote sensing to obtain large-scale images of the disaster-stricken area, and initially understand the overall damage status of power facilities and the distribution of dangerous areas. Dispatch multiple drones equipped with high-definition cameras and lidar equipment to conduct comprehensive video acquisition and three-dimensional terrain scanning of the disaster area, focusing on the damage details of power equipment, such as the tilt angle of towers, the appearance of equipment damage, etc., and arrange professional personnel to carry out rescue work. Personnel carry detection equipment and conduct on-site inspections of some power equipment in a safe area to obtain data such as changes in the equipment's electrical performance parameters. In addition, based on feedback from on-site rescue personnel, dangerous areas are marked, such as areas with leakage risks and areas that may cause secondary dangers due to building collapse. A pre-disaster / post-disaster data mapping algorithm based on feature matching and spatial analysis is used. The geographic location, unique geometric shape, and specific identification of power equipment are used as feature points. By comparing the location of power facilities in geographic space before the disaster with the location changes in post-disaster images and scan data, combined with the changes in the geometric shape of the equipment, an accurate mapping relationship between pre-disaster and post-disaster data is established, achieving deep data fusion. The algorithm formula is: ,in, Indicates the mapping matching degree between pre-disaster and post-disaster data. The smaller the value, the higher the matching degree. is the number of data features, No. The weight of the feature, This is the pre-disaster data eigenvalues, This is the first post-disaster data eigenvalues, It is a supplementary item used to avoid the situation where the denominator is zero. For example, the damage extent and specific damaged parts of the tower can be determined by matching the pre-disaster coordinates and 3D model of the tower with the post-disaster LiDAR scanning data.

[0030] For remote sensing images and drone videos acquired after the disaster, we imported the image and video data, converted their formats, and standardized them. We used the histogram equalization method to enhance image contrast and filter out noise, and extracted spectral, texture, and geometric features. We used the threshold, region, and machine learning-based segmentation method to divide the image into different regions. Through morphological operations and region merging and screening, we optimized the segmentation results. We combined the machine learning recognition model for dangerous areas, determined the exact location and range of dangerous areas based on on-site data and image features, and used the following model formula: ,in, Is the loss function, which is used to measure the difference between the model prediction result and the actual label. is the number of training samples, It is The true label of the sample, 0 represents a non-dangerous area, 1 represents a dangerous area, The model is The probability that a sample is predicted to be a dangerous area, is the parameter vector of the model, is the number of model parameters, and are the regularization coefficients, and the model parameters are adjusted by minimizing the loss function. , so that the model predicts the probability of dangerous areas As close to the true label as possible ,at the same time, and It is used to prevent model overfitting and improve the generalization ability of the model, so as to accurately identify the location, scope and hazard level of dangerous areas, such as dangerous ruins areas formed by building collapse, exposed wire areas that may leak electricity, etc., and classify them according to the degree of danger.

[0031] Based on the topology of the urban power system and the importance of power supply, we identify key equipment for restoring urban power supply, such as core transformers in large substations, key node towers of high-voltage transmission lines connecting different areas, and distribution equipment that supplies power to important public facilities (such as hospitals and fire departments). We assign high priority to the corresponding shooting waypoints of these equipment and follow the strategies of "prioritizing shooting of key equipment" and "low-altitude multi-angle shooting". Under the premise of fully considering avoiding dangerous areas, we use an intelligent path planning algorithm to generate shooting waypoint plans. The algorithm formula is: ,in, is a node The estimated total cost of From the starting point to the node The actual flight cost, is a node Heuristic estimate of the straight-line distance to the target waypoint, is a node The surrounding danger level assessment value is calculated based on the distribution and intensity of the danger zone. and They are the weight coefficients of distance heuristic and degree of danger, respectively. For example, for a damaged core transformer in a large substation, multiple low-altitude waypoints at different heights and angles are planned to ensure that all key parts of the transformer can be photographed in all directions and detailed damage information can be obtained.

[0032] Set rescue progress indicators, such as counting the number of repaired towers, the number of power lines restored, the number of substations inspected, etc., and calculate the corresponding repair progress ratio and inspection completion ratio. The emergency repair demand indicator includes the type and number of newly discovered damaged equipment, special emergency repair needs caused by the earthquake (such as special correction needs for equipment foundation displacement caused by the earthquake, special positioning and repair needs for damaged underground cables, etc.). For example, if a batch of towers with foundation displacement caused by the earthquake are newly discovered, this will be regarded as an important change in the emergency repair demand indicator. When the rescue progress indicator or the emergency repair demand indicator reaches the preset threshold, such as the repair progress is lower than expected for several consecutive hours, or the number of newly discovered severely damaged equipment exceeds a certain proportion of the total number of equipment before the disaster, the waypoint dynamic adjustment mechanism is triggered. By re-evaluating the latest status of key equipment and the dynamic changes in the danger zone (such as the expansion or reduction of the danger zone due to rescue operations), the waypoint adjustment range formula is used to determine the adjustment range. The adjustment range is: ,in, is the adjustment range of waypoint planning, and They are the influence coefficients of changes in rescue progress and power repair needs on the waypoint adjustment range. Based on the calculation results, the waypoint planning is adjusted in a timely and accurate manner to ensure that drone photography can always closely match the actual needs of power repair.

[0033] A communication method combining 5G and wireless private networks is used to ensure stable and high-speed transmission of waypoint planning information and captured image data between the drone and the system in complex urban disaster environments. The transmitted data is encrypted to ensure the security and integrity of the data during transmission. During the data transmission process, the transmission reliability is monitored in real time using the data transmission reliability monitoring formula. The reliability monitoring formula is as follows: ,in, is the reliability score of data transmission, The number of data frames successfully transmitted, is the total number of data frames sent, is the weight of the transmission success rate, is the actual data transmission time, is the expected data transmission time, The weight of transmission time compliance is used to adjust the transmission strategy in a timely manner according to the monitoring results, such as optimizing data encoding methods, switching communication frequency bands, etc., to ensure that the data arrives complete and on time.

[0034] Example 2: Continuous heavy rain caused a debris flow disaster in a mountainous area. A large amount of debris flowed down, burying and destroying some power facilities, resulting in power outage in the area. Some towers were knocked down, cables were buried, and the substation area was also affected by the debris flow, causing varying degrees of damage to equipment.

[0035] Quickly retrieve comprehensive information on the power facilities in the mountainous area before the disaster from the power company's resource database, including the detailed direction of the power lines (taking into account the tortuous changes in the mountainous terrain), the specific location and basic design parameters of each tower, the layout of the substation, and the connection relationship of the internal equipment, etc. At the same time, obtain the topographic data of the area and the topological model of the power system. Use high-resolution satellite remote sensing imagery to obtain an overall image of the disaster-stricken area, focusing on the coverage of the mudslide and the visibility of the power facilities. Send drones with terrain adaptability, equipped with optical and radar sensors, to conduct low-altitude flight photography of the disaster-stricken area, obtain detailed video data of the power facilities, and pay special attention to the location and appearance of equipment buried or impacted by the mudslide. Arrange a survey team composed of professional geological and power technicians to conduct a survey in While ensuring safety, we go deep into the disaster site and use professional geological exploration equipment and power detection tools to collect data such as the thickness of the mudslide, the depth of the buried power equipment, and the damage to the exposed parts of the equipment. We also mark dangerous areas, such as areas where mudslides may occur again, areas where the mountain is unstable due to the impact of mudslides, etc. We use pre-disaster / post-disaster data mapping algorithms, and use the geographic coordinates of the power equipment, the relative position relationship between the equipment and the surrounding terrain, and the characteristic identification of the equipment as feature values. Through complex spatial analysis and data matching, we establish an accurate mapping relationship between pre-disaster and post-disaster data, and integrate the two data. For example, by comparing the position relationship between the tower and the surrounding terrain before the disaster and the terrain changes and tower position changes in the post-disaster images, we can determine the damage and buried location of the tower.

[0036] Post-disaster remote sensing images and drone-captured videos are preprocessed. Specifically, image and video data are imported, converted, and standardized. Histogram equalization is used to enhance image contrast and filter out noise, and spectral, texture, and geometric features are extracted. Segmentation methods based on thresholds, regions, and machine learning are used to divide the image into different regions. Morphological operations and region merging and screening are used to optimize the segmentation results. At the same time, a machine learning identification model for dangerous areas is used, combined with geological survey data and image features, to determine the location and scope of dangerous areas such as potential debris flow and landslide areas and areas where rockfall may occur due to loose mountain, and to assess the level of danger.

[0037] Based on the topological structure and operating characteristics of the power system in mountainous areas, equipment that is critical to restoring power supply in mountainous areas is identified, such as the main equipment of the substation located upstream of the mudslide and controlling the power supply to multiple areas, and the key transmission line towers connecting different mountain villages. These equipment are given high priority. Combined with the flight performance limitations of drones in mountainous areas (for example, due to the influence of mountain airflows, the flight altitude and speed need to be reasonably adjusted), according to the "prioritized shooting of key equipment" and "low-altitude multi-angle shooting" strategies, on the basis of avoiding dangerous areas, the path planning algorithm (taking into account the undulating mountain terrain and obstacles) is used to plan shooting waypoints. For example, for the tower partially buried by the mudslide, waypoints are planned from different angles and heights to clearly capture the buried situation of the tower and the damaged details of the exposed part, and the waypoint path planning formula is used. Generate waypoints where Consider the actual distance and time the drone flies in mountainous areas (due to the influence of terrain, the calculation of flight distance and time needs to take into account factors such as climbing and avoiding obstacles). For nodes Heuristic estimate of the straight-line distance to the target waypoint, For nodes The surrounding danger level assessment value (combined with the distribution of dangerous areas in mountainous areas and geological stability assessment), and They are the weight coefficients of distance heuristic and danger degree, respectively, and are dynamically adjusted according to the actual situation of mountain rescue.

[0038] Set rescue progress indicators, such as counting the number of poles and towers in the area covered by the cleared debris flow, the length of buried cables that have been repaired, the number of mountain villages where power supply has been restored, etc., and calculate the corresponding rescue completion ratio. The emergency repair demand indicators include the number of newly discovered poles and towers with damaged foundations due to the impact of debris flow, the number of excavation and cleaning equipment required, and the demand for emergency repair materials for the special terrain of mountainous areas. For example, if a large number of poles and towers are newly discovered to be loose due to debris flow erosion, this will significantly affect the emergency repair demand indicators. When the rescue progress indicators or emergency repair demand indicators change significantly, such as the expansion of the debris flow burial area, resulting in more than the preset number of poles and towers that cannot be checked in time, the waypoint dynamic adjustment is triggered. The dynamic adjustment trigger condition formula is used to determine whether adjustment is needed, and then the adjustment amount is determined according to the waypoint adjustment amplitude formula. The waypoints are replanned to ensure that drone photography can meet the special needs of mountain power emergency repairs in a timely manner.

[0039] Given the complexity of the communication environment in mountainous areas, a combination of 4G communication and shortwave communication is adopted to ensure the reliability of data transmission in disaster-stricken mountainous environments. The transmitted data is encrypted to ensure data security, and the data transmission status is monitored in real time through the data transmission reliability assessment formula. According to the characteristics of signal changes in mountainous areas, communication parameters such as power and frequency band are adjusted in a timely manner to ensure that waypoint planning information is accurately transmitted to the drone and the captured image data is transmitted back to the system in full to support the smooth progress of damage assessment and emergency repair work of mountainous power facilities.

[0040] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as above in terms of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can, without departing from the scope of the technical solution of the present invention, make some changes or modifications to equivalent embodiments using the technical contents disclosed above. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. An automatic waypoint planning system for electric drone photography, characterized in that: The system includes the following components: data collection and integration module, data analysis and intelligent identification module, waypoint intelligent planning module, dynamic adjustment decision module and communication and data transmission module; the data collection and integration module is responsible for the collection and integration of pre-disaster and post-disaster data, and establishes the mapping relationship between pre-disaster and post-disaster data, minimizes the error through iterative optimization, fuses the mapped data, builds a unified structure and stores it in a database management system; the data analysis and intelligent identification module uses image segmentation technology to pre-process post-disaster remote sensing images and videos, identifies the locations of collapsed buildings and damaged power equipment, collects a large amount of dangerous area sample data and annotates and classifies them, builds a machine learning classification model to identify and classify dangerous areas, and optimizes the model through cross-validation and evaluation indicators; the waypoint intelligent planning module determines the list of key equipment according to the topology and operation rules of the power system and assigns high priority, prioritizes the planning of shooting waypoints for them, and determines the low-altitude flight and multi-angle shooting range according to the performance and shooting requirements of the drone, designs the path to obtain a comprehensive image, and generates initial waypoints in combination with dangerous area information and shooting strategies; The dynamic adjustment decision module sets rescue progress and emergency repair demand indicators, triggers the adjustment mechanism when the indicators change, obtains and analyzes data by establishing a real-time communication interface with the emergency repair command center and on-site personnel, determines the adjustment range of the waypoints, and dynamically adjusts the initial waypoint planning; the communication and data transmission module performs two-way communication through a wireless communication protocol, and formulates a unified transmission format to process waypoint planning and image data, while encrypting the data, establishing a communication link monitoring mechanism to detect transmission reliability in real time, and automatically switching or reconnecting in the event of an abnormality.

2. The automatic planning system for shooting waypoints of electric drones according to claim 1 is characterized in that: The data collection and integration module is responsible for the collection and integration of pre-disaster and post-disaster data. Before the disaster, the location information of power facilities is obtained from the power enterprise resource management system and GIS, and the structural diagram is extracted from the design database. After the disaster, the satellite remote sensing platform is used to collect remote sensing images, and the on-site personnel are organized to collect videos with mobile terminals. The mapping relationship between pre-disaster and post-disaster data is established. The matching degree formula of the mapping relationship is: ,in, It indicates the mapping matching degree between pre-disaster and post-disaster data. The smaller the value, the higher the matching degree. is the number of data features, No. The weight of the feature, This is the pre-disaster data. eigenvalues, This is the first post-disaster data eigenvalues, is a supplementary term to avoid the denominator being zero.

3. The automatic waypoint planning system for electric drone photography according to claim 1 is characterized in that: The data analysis and intelligent recognition module uses image segmentation technology to pre-process post-disaster remote sensing images and videos. Specifically, it imports image and video data and converts and standardizes their formats, uses the histogram equalization method to enhance image contrast and filter out noise, and extracts spectral, texture and geometric features. It uses a segmentation method based on thresholds, regions and machine learning to divide the image into different regions, and optimizes the segmentation results through morphological operations and region merging and screening processing.

4. The automatic waypoint planning system for electric drone photography according to claim 1 is characterized in that: The data analysis and intelligent identification module constructs a machine learning classification model to identify and classify dangerous areas, and its model formula is: ,in, is the loss function, which is used to measure the difference between the model prediction result and the actual label. is the number of training samples, It is The true label of the sample, 0 represents a non-dangerous area, 1 represents a dangerous area, The model is The probability that a sample is predicted to be a dangerous area, is the parameter vector of the model, is the number of model parameters, yes Regularization term, yes Regularization term, and The regularization terms are and coefficient.

5. The automatic waypoint planning system for electric drone photography according to claim 1 is characterized in that: The waypoint intelligent planning module determines a list of key equipment according to the topology and operation rules of the power system and assigns high priority to it, and prioritizes the planning of shooting waypoints for it. Specifically, the topological connection relationship data and operation rule data of the power system are obtained. Based on the data, the power system is abstracted into a topological graph. By analyzing the connectivity of the graph, the nodes and connection paths whose faults will cause large-scale power outages or affect important power supply areas are found. At the same time, the power distribution during normal operation of the power system is analyzed, and the equipment and lines that carry large power transmission tasks are determined. According to the rated parameters of the equipment, large-capacity generators and main transformer equipment are screened out. In combination with the load rate and the protection requirements of important users, the key equipment for long-term high-load operation and power supply to important users is determined, a list of key equipment is formed, and a high-priority mark is assigned to the key equipment in the list. The shooting area is determined according to the equipment location coordinates and the drone shooting angle and resolution requirements. The intelligent path planning algorithm is used to generate waypoints that meet the "low-altitude multi-angle shooting" strategy in the shooting area. When generating waypoints, the drone flight performance limitations and dangerous area information are considered. According to the priority order of key equipment, shooting waypoints are planned for high-priority key equipment in priority order.

6. The automatic waypoint planning system for electric drone photography according to claim 5 is characterized in that: The waypoint intelligent planning module uses an intelligent path planning algorithm to generate waypoints that meet the "low-altitude multi-angle shooting" strategy in the shooting area. The algorithm formula is: ,in, Is a node The estimated total cost of From the starting point to the node The actual flight cost, Is a node Heuristic estimate of the straight-line distance to the target waypoint, Is a node The surrounding danger level assessment value is calculated based on the distribution and intensity of the danger zone. and are the weight coefficients of distance heuristic and danger level respectively.

7. The automatic waypoint planning system for electric drone photography according to claim 1 is characterized in that: The dynamic adjustment decision module sets rescue progress and repair demand indicators. For the rescue progress indicator, the equipment repair situation is obtained from the power repair command center, the number of repairs is counted by type and compared with the total number before the disaster, the overall repair progress ratio is calculated by weight, the power supply area is divided according to the power topology and GIS data, the number of restored power supply areas is counted with the help of the power dispatching system, the power supply restoration ratio is obtained, and the repair team is equipped with positioning equipment, and its advancement position is determined in combination with the task point, and the task completion ratio is calculated to comprehensively measure the overall advancement index. For the power repair demand indicator, the newly damaged equipment information is collected through real-time image recognition and on-site feedback, and classified and counted by type and degree of damage, and special repair needs are collected and quantified through communication with all parties. The dynamic information of the dangerous area is obtained from the data analysis and intelligent identification module, and its scope and level changes are analyzed and quantified. The historical data of disaster repair is analyzed to set the initial threshold for each indicator, and the threshold is dynamically optimized according to the indicator data in the actual rescue and the adjustment effect of the waypoint planning. For the rescue progress indicator, the adjustment formula is: ,in, is the change in rescue progress, and are the rescue progress indicators at the current and previous moments, is the trigger threshold of the change in rescue progress. For the power repair demand index, the adjustment formula is: ,in, is the change in power repair demand, and are the power repair demand indicators at the current and previous moments, It is the trigger threshold for changes in power emergency repair demand.

8. The automatic waypoint planning system for electric drone photography according to claim 7 is characterized in that: The dynamic adjustment decision module acquires and analyzes data by establishing a real-time communication interface with the emergency repair command center and on-site personnel to determine the adjustment range of the waypoint. The adjustment range is: ,in, is the adjustment range of waypoint planning, and They are the impact coefficients of changes in rescue progress and power repair demand on the waypoint adjustment range.

9. The automatic waypoint planning system for electric drone photography according to claim 1 is characterized in that: The communication and data transmission module establishes a communication link monitoring mechanism to detect transmission reliability in real time, and its reliability monitoring formula is: ,in, is the reliability score of data transmission, The number of successfully transmitted data frames, is the total number of data frames sent, is the weight of the transmission success rate, is the actual data transmission time, is the expected data transfer time, The weight of the transmission time compliance.

Images