Dynamic external damage prevention monitoring and early warning system and method for power equipment

Abstract

The invention relates to the technical field of line external damage prevention, in particular to a dynamic external damage prevention monitoring and early warning system and method for power equipment. The system comprises an unmanned aerial vehicle which carries a camera and is used for monitoring the power equipment and the surrounding environment in real time, and the unmanned aerial vehicle sends monitored data to an image processing module; the data processing module is used for receiving monitoring data of the unmanned aerial vehicle and preprocessing the monitoring data; the identification and tracking module is used for carrying out multi-target identification and real-time tracking on the preprocessed data; the alarm module is used for giving an alarm if the recognition and tracking module detects that the power equipment is threatened by the external damage target; and the control module is used for adjusting the flight height of the unmanned aerial vehicle and the shooting angle of the camera. The external damage target can be monitored, recognized and early warned in real time, so that countermeasures are taken in time, and accidents are prevented from occurring or expanding.

Description

Technical Field

[0001] The present invention relates to the technical field of line external damage prevention, and in particular to a dynamic external damage prevention monitoring and early warning system and method for electric power equipment. Background Art

[0002] Traditional power transmission channel inspections mainly rely on manual inspections, which have problems such as low work efficiency, high human resource costs, and insufficient blind spot coverage. Some power transmission channels are monitored by fixed cameras, but they are limited to fixed areas and cannot achieve comprehensive monitoring of the entire channel.

[0003] In recent years, drone technology has developed rapidly, and its application scope covers various fields. In the safety management of power transmission channels, the application of drones can greatly improve monitoring efficiency and coverage, while reducing the consumption of human resources. However, existing products can often only detect and identify specific types of external damage targets, and have limited recognition capabilities for complex scenes or new targets. For example, the drone-based power equipment anti-external damage warning method, system and device disclosed in publication number: CN118196976A mainly monitor engineering vehicles, and cannot monitor the environment around the power equipment more comprehensively. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and provide a dynamic anti-external damage monitoring and early warning system and method for electric power equipment.

[0005] The technical solution adopted by the present invention to solve the technical problem is: a dynamic anti-external damage monitoring and early warning system for power equipment includes:

[0006] Drone: Equipped with a camera, it is used to monitor the power equipment and its surrounding environment in real time. The drone sends the monitoring data to the image processing module. Specifically, the power equipment and its surrounding environment are monitored in real time from the air by the drone, and the video stream or image data is obtained through the camera, and then the data is sent to the data processing module.

[0007] Data processing module: receives the monitoring data from the drone and performs pre-processing;

[0008] Identification and tracking module: used to perform multi-target identification and real-time tracking on pre-processed data;

[0009] Alarm module: If the identification and tracking module detects that the power equipment is threatened by external targets, an alarm will be issued;

[0010] Control module: used to adjust the flight altitude of the drone and the shooting angle of the camera.

[0011] The identification and tracking module includes

[0012] Identification unit: used to identify and locate external targets of power equipment and its surroundings;

[0013] Tracking unit: used to realize multi-target tracking, target matching and proximity detection.

[0014] The identification unit identifies external destruction targets of the power equipment and its surroundings, assigns a unique identifier to each external destruction target, and records the location and operating status of the external destruction target, and then sends the external destruction target to the tracking unit.

[0015] The tracking unit tracks multiple targets: based on the motion model and historical trajectory of the external target, predicts the future motion trajectory of the external target and updates it in real time;

[0016] For target matching: compare the observed new external target with the known external target. If the similarity between the features or motion trajectory of the known external target and the new external target is within the set range, associate them to determine the matching relationship between them. According to the matching results, update the status of each external target to achieve continuous tracking and identification of the target.

[0017] For proximity detection: a safety range is set based on the power equipment to determine whether the external damage target is within the safety range or is approaching the safety range. If so, an alarm is issued through the alarm module.

[0018] It also includes a lighting correction module, which is used to perform lighting correction on the image to improve the accuracy of target detection and tracking.

[0019] The illumination correction module comprises

[0020] Illumination angle calculation unit: used to calculate the angle θ between the incident light and the surface normal;

[0021] Incident light intensity acquisition unit: used to acquire incident light intensity I0;

[0022] Reflected light intensity calculation unit: used to calculate the reflected light intensity I of each pixel in the image;

[0023] Light correction unit: used to calculate the reflected light intensity I that should be corrected corrected , perform image correction.

[0024] The camera is provided with an automatic exposure control module, which is connected to the illumination correction unit and is used to receive the corrected pixel brightness I corrected, automatically adjust the exposure parameters to improve the clarity of the image when shooting. According to the Lambertian reflection model, the light intensity is obtained, and the exposure parameters of the camera are automatically adjusted according to the ambient light conditions to ensure that the captured image has moderate brightness and can clearly display the target. The Lambertian reflection model can be used to help the system judge the lighting conditions, thereby realizing automatic exposure control.

[0025] The recognition and tracking module performs target recognition and tracking based on the corrected image. The corrected image has more uniform brightness, which helps to improve the detection accuracy of the YOLO model and the state estimation accuracy of the particle filter.

[0026] The method applied to the above-mentioned dynamic anti-external damage monitoring and early warning system for electric power equipment comprises the following steps:

[0027] Step 1: Obtain image information of power equipment and its surrounding environment based on the drone, and preprocess it through the data processing module; the drone's camera can capture video streams or take images at an appropriate frequency. The acquired video streams or images need to be preprocessed to improve the efficiency and accuracy of target detection. Preprocessing includes operations such as image scaling, cropping, and brightness adjustment to meet the input requirements of subsequent target detection algorithms.

[0028] Step 2: Based on the pre-processed image information, identify the external damage targets of the power equipment and its surroundings, and record the location and operating status of the external damage targets;

[0029] Step 3: Track the external destruction target in real time, predict the future movement trajectory of the external destruction target, and update it in time;

[0030] Step 4: Determine whether the external destruction target is a floating object. If so, issue an alarm message; otherwise, determine whether the external destruction target has entered the safety range of the power equipment or whether there is a trend of entering the safety range. If so, issue an alarm message, otherwise continue to track the external destruction target in real time. If the external destruction target stops moving before entering the safety range, the alarm can be stopped; if the power equipment has been destroyed, an alarm is issued while sending a maintenance signal to notify the maintenance personnel to repair it in time.

[0031] The following steps are also included:

[0032] Step 5: Obtain a new external image, and perform target recognition and tracking based on the new external image. The clarity of the new image is greater than that of the original image.

[0033] The new external image is obtained as follows:

[0034] Lowering the cruising altitude of the drone, controlling the drone to approach the external destruction target, adjusting the camera shooting angle, and obtaining image information of the external destruction target;

[0035] and / or

[0036] Calculate the reflected light intensity I that should be corrected corrected , obtain the corrected image, perform target recognition and tracking based on the corrected image, and automatically adjust the camera exposure parameters to shoot the externally broken target to obtain the externally broken image.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The system can realize dynamic protection against external damage by combining drones with multiple algorithms, effectively improving the safety and reliability of key facilities such as power equipment (such as transmission lines). The system can monitor, identify and warn external damage targets in real time, so as to take timely response measures to prevent accidents from happening or expanding.

[0039] 2. The system uses lightweight artificial intelligence technology and one algorithm to achieve multiple functions, which enables the system to have efficient target recognition and detection capabilities while reducing overall energy consumption. Compared with traditional manual inspections, the system can achieve functions such as autonomous identification, dynamic route planning, and pull flow identification, which greatly improves safety management efficiency while reducing human resource consumption and inspection costs;

[0040] 3. This application monitors and identifies external targets, abnormal situations, etc. in real time, can promptly discover potential safety hazards, and help prevent power equipment from being destroyed or damaged, thereby improving the safety of power transmission channels. Through the automated monitoring and identification system, the monitoring efficiency can be greatly improved, and real-time monitoring of a large area can be achieved, thereby reducing the consumption of human resources and monitoring costs, and improving work efficiency.

[0041] 4. YOLO is used to realize anti-external damage target detection and recognition, fireworks detection, and engineering vehicle detection functions. The particle filter realizes multi-target tracking, target matching, and proximity detection. The Lambertian reflection model performs image illumination correction to improve image clarity. Based on algorithm fusion, energy consumption is reduced.

[0042] 5. The application of the illumination correction module improves the image quality, especially when the illumination conditions vary greatly, ensuring the accuracy of target detection and tracking; the automatic exposure control module automatically adjusts the camera parameters according to the illumination conditions, enhancing the adaptability and robustness of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] The dynamic anti-external damage monitoring and early warning system for power equipment includes

[0046] Drone: Equipped with a camera, it is used to monitor the power equipment and its surrounding environment in real time. The drone sends the monitoring data to the image processing module. Specifically, the power equipment and its surrounding environment are monitored in real time from the air by the drone, and the video stream or image data is obtained through the camera, and then the data is sent to the data processing module.

[0047] Data processing module: receives the monitoring data from the drone and performs pre-processing;

[0048] Identification and tracking module: used to perform multi-target identification and real-time tracking on the pre-processed data; the identification and tracking module includes an identification unit and a tracking unit;

[0049] Among them, the recognition unit is used to identify and locate external targets of power equipment and its surroundings; the recognition unit of this application adopts a trained YOLO model. The YOLO model is a deep neural network that can simultaneously realize target detection and positioning, and can predict the category and position of all targets in one forward propagation. The preprocessed video stream or image is sent to the trained YOLO model for target detection.

[0050] If the target is electrical equipment, the external destruction targets may include the following categories: cranes, tower cranes, cement pump trucks, bulldozers, forklifts, excavators, dump trucks, other construction machinery, foreign objects, wildfires, smoke and other equipment and objects or situations that may pose a threat to electrical equipment.

[0051] The YOLO model can identify various target categories in the input image, such as construction vehicles, cranes, tower cranes, etc. Each target category has a corresponding category label. By using a rectangular bounding box to represent the location of the target in the image and giving the coordinate information of the bounding box, the location of the target can be output.

[0052] The YOLO model can detect multiple targets at the same time, so it can effectively deal with various possible external targets around the transmission channel, such as construction vehicles, cranes, etc. In one detection, multiple targets can be identified and their locations and categories can be output together.

[0053] YOLO also performs engineering vehicle recognition here, and performs real-time analysis on video streams or images to detect engineering vehicles. Engineering vehicles mainly include excavators, bulldozers, trucks, mixer trucks, cranes, forklifts, cranes, etc. Among them, excavators, bulldozers, cranes, cranes, etc. are key monitoring objects.

[0054] In addition, YOLO can also monitor and identify power transmission channels and surrounding environments in real time, and promptly detect and warn of safety hazards such as fireworks. When the external target is fireworks, the alarm module can immediately sound an alarm, and the system takes corresponding response measures, such as notifying relevant personnel and launching emergency plans, to ensure safety. Unlike detecting other external targets, the system will sound an alarm immediately when fireworks are detected, rather than only when it detects that the target poses a safety threat to power equipment.

[0055] For areas prone to wildfires such as forests, drones can be equipped with the DJI Matrice3TD model in DJI Airport 2. This model is equipped with an infrared camera and combined with the YOLO model, which can more clearly detect fire points and further improve the recognition rate of wildfires. Under normal circumstances, ordinary models can be configured to reduce overall costs.

[0056] The tracking unit is used to realize multi-target tracking, target matching and proximity detection. In this application, the tracking unit is a particle filter.

[0057] The identification unit identifies external destruction targets of the power equipment and its surroundings, assigns a unique identifier to each external destruction target, and records the location and operating status (position, speed, etc.) of the external destruction target, and then sends the external destruction target to the tracking unit.

[0058] The tracking unit uses a particle filter. For multi-target tracking: based on the motion model and historical trajectory of the external target, the future motion trajectory of the external target is predicted and updated in real time; in the multi-target tracking process, there may be target overlap, occlusion, etc., and the targets need to be associated, that is, the relationship between the targets is determined to ensure that each target is tracked correctly. At the same time, data association is also required to associate the tracking information of each target with other sensor data (such as infrared sensors, etc.) to improve the accuracy and robustness of tracking.

[0059] For target matching: compare the observed new external target with the known external target. If the similarity between the features or motion trajectory of the known external target and the new external target is within the set range, associate them to determine the matching relationship between them. According to the matching results, update the status of each external target to achieve continuous tracking and identification of the target.

[0060] Specifically, first, the particle filter uses the observed data to update the state of the target. This is achieved by comparing the predicted state of the target with the actual observed data and calculating the weight of each particle. The weight reflects the degree of fit between the particle and the observed data, that is, the possible position of the target at the current moment.

[0061] Then, target association is performed using the known targets and their state estimates, as well as the observed new targets. The observed targets are matched by comparing their distance or similarity with the known targets, and the target states are updated.

[0062] Once the target association is completed, the target can be identified and matched. This means determining the correspondence between the observed target and the known targets and associating them. The observed target is compared with the known targets with similar features or trajectories to determine the matching relationship between them.

[0063] Finally, based on the target matching results, the state estimate of each target is updated. This includes updating the target's position, speed and other information, and adjusting and correcting the state based on the new observation data. This allows for continuous tracking and identification of the target.

[0064] For proximity detection: a safety range is set based on the power equipment to determine whether the external damage target is within the safety range or is approaching the safety range. If so, an alarm is issued through the alarm module.

[0065] After the target is matched, a proximity detection is performed to confirm whether the target will have an impact on the power equipment.

[0066] First, the particle filter uses the historical motion data and motion model of the target to predict the position and trajectory of the target at future moments. This prediction process takes into account the dynamic characteristics of the target such as speed and acceleration, as well as environmental factors such as wind speed and terrain. A safety range or warning area is set in the system (here the warning area for power equipment can be set to 20 meters to ensure the safety of power equipment) to indicate the area or facility that the system is concerned about. This safety range can be set according to specific application scenarios and requirements, such as no-fly zones around transmission channels, no-entry zones at airports, etc.

[0067] The predicted position of the target is compared with the safety range set by the system to determine whether the target is close to the safety range. If the predicted position of the target is within the safety range, or is close enough to the boundary of the safety range, it can be considered that the target has approached the area or facility of concern to the system. Once the target is found to be close to the safety range, the system can take corresponding early warning and protective measures. At the same time, the drone hovers above the target to monitor the target's movements.

[0068] Alarm module: If the identification and tracking module detects that the power equipment has the threat of external damage targets, an alarm will be issued; the alarm module of this application includes the integration of the loudspeaker and the RTK data interface set on the drone: a technical solution that combines the loudspeaker with the RTK (real-time motion positioning) data interface. By connecting the RTK positioning system with the loudspeaker, real-time transmission and processing of positioning data is achieved, and it is used for precise shouting functions. Through this solution, the loudspeaker can make directional shouting to specific areas or objects based on the precise location information provided by the RTK positioning system, improving the shouting effect and coverage.

[0069] The megaphone can be used to send warning and alert sounds to the target, informing it that it has been detected and is under surveillance. This can have a deterrent effect, making the target aware that it is being detected and may reduce its potential inappropriate behavior. The megaphone can also be used to send instructions and guidance information to the target. When the drone approaches the target, the megaphone can send instructions to the target, telling it how to cooperate with the monitoring or how to act to ensure the safety and smooth progress of the flight mission.

[0070] For example, you can use the megaphone to announce "Warning! You have entered a restricted area, please leave immediately!", "Caution! Do not approach electrical equipment to avoid danger!", and when large equipment is too close, shout "Please keep a safe distance to avoid collision!"

[0071] If the target does not listen to dissuasion and has caused damage to the power equipment, on the one hand, the maintenance personnel will be notified to carry out timely maintenance. On the other hand, the drone will continue to track the target that has damaged the power equipment and broadcast its location in real time to prevent the target from escaping and facilitate subsequent accountability for the target.

[0072] Control module: used to adjust the flight altitude of the drone and the shooting angle of the camera.

[0073] It also includes an illumination correction module, which is used to perform illumination correction on the image to improve the accuracy of target detection and tracking. The illumination correction module of this application uses the Lambertian reflection model to improve the accuracy and stability of target detection and tracking. The Lambertian reflection model is a simplified reflection model that assumes that the surface reflects light with the same brightness in all directions and is suitable for illumination simulation of rough surfaces.

[0074] Furthermore, the illumination correction module includes

[0075] Illumination angle calculation unit: used to calculate the angle θ between the incident light and the surface normal;

[0076] Incident light intensity acquisition unit: used to acquire incident light intensity I0;

[0077] Reflected light intensity calculation unit: used to calculate the reflected light intensity I of each pixel in the image;

[0078] Light correction unit: used to calculate the reflected light intensity I that should be corrected corrected , perform image correction.

[0079] The basic calculation formula for reflected light intensity I is:

[0080] I = I0·cos(θ);

[0081] In practical applications, the surface reflectivity ρ also needs to be considered, and the calculation formula is:

[0082] I = ρI0·cos(θ);

[0083] In actual scenes, lighting conditions may change due to factors such as time, weather, and geographic location, resulting in uneven brightness of the target in the image. Using the Lambertian reflection model, the image can be corrected for lighting, making the target clearer under different lighting conditions.

[0084] Assuming that the incident light comes from a single light source and its direction is known, the reflected light intensity correction value I corrected The calculation steps are:

[0085] Calculate the angle θ between the incident light and the surface normal (i.e. the angle of the incident light): Let the direction of the incident light be vector L, and the direction of the normal be vector N, then the angle of incidence θ is:

[0086]

[0087] Obtain incident light intensity I0: It is usually known or directly measured by existing instruments (such as optical power meter, spectroradiometer, etc.).

[0088] Then calculate the reflected light intensity I of each pixel point, and the calculation formula is:

[0089] I=ρI o ·cos(θ);

[0090] The reflected light intensity I should be corrected corrected , the calculation formula is:

[0091]

[0092] In the formula, I original is the pixel brightness of the original image, and the reflected light intensity I should be corrected corrected The brightness of each pixel in the image is adjusted to the corresponding pixel brightness to be corrected to reduce the impact of uneven lighting.

[0093] Images captured by drone cameras under different lighting conditions may have uneven brightness. By applying the Lambertian reflection model, the brightness of different areas can be adjusted according to the lighting conditions in the image to obtain a more balanced and realistic image effect.

[0094] The camera is provided with an automatic exposure control module, which is connected to the illumination correction unit and is used to receive the corrected pixel brightness I corrected , automatically adjust the exposure parameters to improve the clarity of the image when shooting. According to the Lambertian reflection model, the light intensity is obtained, and the exposure parameters of the camera are automatically adjusted according to the ambient light conditions to ensure that the captured image has moderate brightness and can clearly display the target. The Lambertian reflection model can be used to help the system judge the lighting conditions, thereby realizing automatic exposure control.

[0095] The recognition and tracking module performs target recognition and tracking based on the corrected image. The corrected image has more uniform brightness, which helps to improve the detection accuracy of the YOLO model and the state estimation accuracy of the particle filter.

[0096] Assume that a drone inspects the same power tower at different time periods and obtains the following images:

[0097] Morning: The sun shines obliquely from the left, the left side of the power tower is bright and the right side is dark.

[0098] Noon: The sun shines vertically, and the power tower is relatively evenly distributed overall.

[0099] In the evening: the sun shines obliquely from the right, the right side of the power tower is bright and the left side is dark.

[0100] Through the above illumination correction steps, the brightness distribution of the three images can be made more uniform, thereby improving the robustness and accuracy of target detection and tracking.

[0101] The surface reflectivity ρ can be measured experimentally, but the power inspection here does not have the ability to conduct experiments, mainly because there is an internal database. During the power inspection process, although there is no ability to experimentally measure the surface reflectivity, an internal database can be used, which contains the reflectivity information of common power equipment materials. Use the YOLO model to detect and identify target devices in the image. For example, the detected target is an excavator. Look up the reflectivity of the main material (metal) of the excavator from the database, assuming that ρ = 0.6. Perform illumination correction based on ρ = 0.6. The following is a reflectivity database table, including common equipment and materials that may cause external damage in power inspections.

[0102] Reflectivity database table

[0103] Device Type Main Materials Reflectivity ρ\rhoρ excavator Metal 0.6 bulldozer Metal 0.6 car Metal 0.6 truck Metal 0.6 tractor Metal 0.6 Plastic pipe plastic 0.4 Plastic Box plastic 0.4 cable plastic 0.4 Porcelain Insulator ceramics 0.5 Ceramic tube ceramics 0.5 Telephone pole Concrete 0.3 Road Obstacles Concrete 0.3 Trees plant 0.2

[0104] Reference Figure 1 , the method applied to the above-mentioned power equipment dynamic anti-external damage monitoring and early warning system comprises the following steps:

[0105] Step 1: Obtain image information of power equipment and its surrounding environment based on the drone, and preprocess it through the data processing module; the camera of the drone can capture video streams or take images at an appropriate frequency. The acquired video streams or images need to be preprocessed to improve the efficiency and accuracy of target detection. Preprocessing includes operations such as image scaling, cropping, and brightness adjustment to meet the input requirements of subsequent target detection algorithms. The preprocessed video streams or images are fed into the trained YOLO model for target detection. The YOLO model is a deep neural network that can simultaneously realize target detection and positioning, and can predict the categories and locations of all targets in one forward propagation.

[0106] Step 2: Based on the pre-processed image information, identify the external damage targets of the power equipment and its surroundings, and record the location and operating status of the external damage targets;

[0107] Step 3: Track the external target in real time, predict the future trajectory of the external target (based on the target's motion model and its historical trajectory), and update it in time;

[0108] Step 4: Determine whether the external destruction target is a floating object (such as smoke, etc.). If so, issue an alarm message; otherwise, determine whether the external destruction target has entered the safety range of the power equipment or whether there is a trend of entering the safety range. If so, issue an alarm message, otherwise continue to track the external destruction target in real time. If the external destruction target stops moving before entering the safety range, the alarm can be stopped; if the power equipment has been destroyed, an inspection signal is issued at the same time as the alarm to notify the maintenance personnel to repair it in time.

[0109] The following steps are also included:

[0110] Step 5: Obtain a new external image, and perform target recognition and tracking based on the new external image. The clarity of the new image is greater than that of the original image.

[0111] The new external image is obtained as follows:

[0112] Lower the cruising altitude of the drone, control the drone to approach the external destruction target, adjust the camera shooting angle, and obtain image information of the external destruction target, thereby capturing or recording clearer images or video streams;

[0113] and / or

[0114] Calculate the reflected light intensity I that should be corrected corrected , obtain the corrected image, perform target recognition and tracking based on the corrected image, and automatically adjust the camera exposure parameters to shoot the externally broken target to obtain the externally broken image, thereby shooting a clearer externally broken image.

[0115] This application sets up an edge computing box on the drone. This application uses NVIDIA's NVIDIA Xavier NX, with an AI performance of 21TOPS. The performance meets the business scenarios and the equipment is stable.

[0116] The edge computing box can process and analyze data at a location close to the source of data generation (i.e., on the drone) instead of sending the data to a remote server for processing. This can reduce data transmission delays and network bandwidth occupancy, and improve data processing efficiency and real-time performance. The core of anti-external damage is to protect power equipment from external damage or destructive events through effective monitoring, identification and preventive measures. For target identification and obstacle avoidance, the main functions involved are anti-external damage target detection and recognition, fireworks detection, engineering vehicle detection, illumination model, multi-target locking and tracking, target matching, and proximity detection. Each of these is implemented using an independent algorithm. For the edge computing box on the drone, if all independent algorithms are used to implement it, the amount of calculation will be very large, which is too much for the edge computing box, and the energy consumption will also be very high. Therefore, this application uses a fusion algorithm, that is, one algorithm implements multiple functions to reduce the overall computing overhead.

[0117] The algorithms used in this application include the YOLO model, the particle filter, and the Lambertian reflection model. The fusion of these three algorithms realizes the functions that originally required multiple algorithms to achieve. Among them, YOLO can realize anti-external damage target detection and recognition, fireworks detection, and engineering vehicle detection functions. The particle filter is used to achieve multi-target tracking, target matching, and proximity detection. The Lambertian reflection model is an illumination model that performs illumination correction on the image to improve image clarity. Through the combination of these three algorithms, the functions that originally required seven algorithms can be achieved.

[0118] In addition, the location of the drone airport is set according to the power lines. At present, compared with the first generation of the airport, the volume and weight of DJI Airport 2 are reduced by 75% and 68% respectively, and two people can carry it, which greatly improves the installation flexibility and reduces the installation cost. Therefore, this application uses DJI Airport 2.

[0119] The airport adopts a high-protection design, and the dust and water resistance of the whole machine reaches IP55 level. The airport integrates sensors such as rain gauge, anemometer, thermometer, etc., which can sense weather changes in real time. Combined with online weather forecasts, DJI Sky2 can issue timely warnings or terminate the original flight mission, effectively reducing flight risks.

[0120] Thanks to the new generation of image recognition technology, the aircraft can more accurately identify the landing signs above the apron. Not only that, the apron adopts a new sliding slope return design to further improve the landing success rate. With a built-in backup battery, if an unexpected power outage occurs, the airport can still work independently for more than 5 hours with the help of the built-in battery.

[0121] Thanks to the airport's high protection design and high reliability, the airport only needs maintenance once every six months, effectively controlling manpower investment costs. When the airport and aircraft encounter mission failures or emergencies, DJI Sky2 will send SMS and email reminders immediately, and operators can accurately troubleshoot the problem based on the guidance information.

[0122] Since the coverage range of DJI Airport 2 is 10 kilometers, the location of the drone airport needs to be set according to the layout of the power equipment. In addition, additional lines can be set according to special circumstances. For example, when a new construction site is started near the power equipment, a drone airport can be set up at the corresponding location to achieve monitoring.

[0123] The above descriptions are only optional embodiments of the present invention, and are not intended to limit the patent scope of the present invention. All equivalent structural changes made using the contents of the present invention's specification and drawings, or directly / indirectly applied in other related technical fields, are included in the patent protection scope of the present invention.

Claims

1. Dynamic anti-external damage monitoring and early warning system for power equipment, characterized by: include Drone: Equipped with a camera, it is used to monitor the power equipment and its surrounding environment in real time. The drone sends the monitoring data to the image processing module; Data processing module: receives the monitoring data from the drone and performs pre-processing; Identification and tracking module: used to perform multi-target identification and real-time tracking on pre-processed data; Alarm module: If the identification and tracking module detects that the power equipment is threatened by external targets, an alarm will be issued; Control module: used to adjust the flight altitude of the drone and the shooting angle of the camera.

2. The dynamic anti-external damage monitoring and early warning system for electric power equipment according to claim 1 is characterized in that: The identification and tracking module includes Identification unit: used to identify and locate external targets of power equipment and its surroundings; Tracking unit: used to realize multi-target tracking, target matching and proximity detection.

3. The dynamic anti-external damage monitoring and early warning system for electric power equipment according to claim 2 is characterized in that: The identification unit identifies external destruction targets of the power equipment and its surroundings, assigns a unique identifier to each external destruction target, and records the location and operating status of the external destruction target, and then sends the external destruction target to the tracking unit.

4. The dynamic anti-external damage monitoring and early warning system for electric power equipment according to claim 3 is characterized in that: The tracking unit tracks multiple targets: based on the motion model and historical trajectory of the external target, predicts the future motion trajectory of the external target and updates it in real time; For target matching: compare the observed new external target with the known external target. If the similarity between the features or motion trajectory of the known external target and the new external target is within the set range, associate them to determine the matching relationship between them. According to the matching results, update the status of each external target to achieve continuous tracking and identification of the target. For proximity detection: a safety range is set based on the power equipment to determine whether the external damage target is within the safety range or is approaching the safety range. If so, an alarm is issued through the alarm module.

5. The dynamic anti-external damage monitoring and early warning system for electric power equipment according to claim 1 is characterized in that: It also includes a lighting correction module, which is used to perform lighting correction on the image to improve the accuracy of target detection and tracking.

6. The dynamic anti-external damage monitoring and early warning system for electric power equipment according to claim 5 is characterized in that: The illumination correction module comprises Illumination angle calculation unit: used to calculate the angle θ between the incident light and the surface normal; Incident light intensity acquisition unit: used to acquire incident light intensity I0; Reflected light intensity calculation unit: used to calculate the reflected light intensity I of each pixel in the image; Light correction unit: used to calculate the reflected light intensity I that should be corrected corrected , perform image correction.

7. The dynamic anti-external damage monitoring and early warning system for electric power equipment according to claim 6 is characterized in that: The camera is provided with an automatic exposure control module, which is connected to the illumination correction unit and is used to receive the corrected pixel brightness I corrected , automatically adjust exposure parameters to improve image clarity when shooting.

8. The dynamic anti-external damage monitoring and early warning system for electric power equipment according to claim 6 is characterized in that: The recognition and tracking module performs target recognition and tracking based on the corrected image.

9. The method applied to the dynamic anti-external damage monitoring and early warning system for power equipment according to any one of claims 1 to 8, characterized in that: The steps include: Step 1: Obtain image information of power equipment and its surrounding environment based on drones, and pre-process it through a data processing module; Step 2: Based on the pre-processed image information, identify the external damage targets of the power equipment and its surroundings, and record the location and operating status of the external damage targets; Step 3: Track the external destruction target in real time, predict the future movement trajectory of the external destruction target, and update it in time; Step 4: Determine whether the external destruction target is a floating object. If so, issue an alarm message; otherwise, determine whether the external destruction target has entered the safe range of the power equipment or whether there is a trend of entering the safe range. If so, issue an alarm message, otherwise continue to track the external destruction target in real time. If the power equipment has been damaged, issue a maintenance signal at the same time as the alarm to notify the maintenance personnel to repair it in time.

10. The method according to claim 9, characterized in that The following steps are also included: Step 5: Obtain a new external image, and perform target recognition and tracking based on the new external image. The clarity of the new image is greater than that of the original image. The new external image is obtained as follows: Lowering the cruising altitude of the drone, controlling the drone to approach the external destruction target, adjusting the camera shooting angle, and acquiring image information of the external destruction target; and / or Calculate the reflected light intensity I corrected , obtain the corrected image, perform target recognition and tracking based on the corrected image, and automatically adjust the camera exposure parameters to shoot the externally broken target to obtain the externally broken image.

Images