Unmanned aerial vehicle inspection method and device for high-risk closed scene and medium

Abstract

The invention discloses an unmanned aerial vehicle inspection method and device for a high-risk closed scene and a medium, and the method comprises the steps: dividing the high-risk closed scene into a plurality of inspection regions, and determining the inspection priorities corresponding to the inspection regions based on historical data; determining a plurality of sub-task packets corresponding to the plurality of inspection areas, and determining a target sub-task packet corresponding to the unmanned aerial vehicle according to the basic information of each unmanned aerial vehicle; predicting a movement track of the intelligent mobile device based on the intelligent mobile device information, and determining a tour inspection path corresponding to the unmanned aerial vehicle in combination with a preset 3D point cloud map corresponding to the high-risk closed scene; and constructing an electric quantity task coupling model to predict an executable task duration of the unmanned aerial vehicle based on the inspection path, and continuing inspection through the replaceable unmanned aerial vehicle after the unmanned aerial vehicle executes the executable task duration.

Description

Technical Field

[0001] This application relates to the technical field of intelligent patrol inspection, and particularly to a drone patrol inspection method, device and medium for high-risk enclosed scenarios. Background Art

[0002] With the acceleration of the industrialization process and the deepening of educational and scientific research activities, the demand for safety monitoring in high-risk enclosed scenarios has shown an explosive growth, such as school chemistry laboratories, underground chemical warehouses, internal pipelines of nuclear facilities, biosafety protection laboratories, etc. High-risk enclosed scenarios have risks such as toxic gas leakage, accumulation of flammable and explosive substances, and radioactive contamination. Narrow passages, dense shelves or complex pipeline structures limit the movement of equipment, and different scenarios may also require low-temperature, high-temperature or high-humidity environments, featuring high environmental risk, spatial limitation and special climate.

[0003] Traditional manual patrol inspection relies on operators to enter high-risk areas wearing heavy protective equipment. However, manual detection has defects such as the risk of toxicity exposure, the limit of extreme environment tolerance, and slow operation and low work efficiency, making it difficult to meet the patrol inspection requirements in high-risk enclosed scenarios. Summary of the Invention

[0004] Embodiments of this application provide a drone patrol inspection method, device and medium for high-risk enclosed scenarios to solve the above technical problems.

[0005] On the one hand, embodiments of this application provide a drone patrol inspection method for high-risk enclosed scenarios, including: Dividing the high-risk enclosed scenario into multiple patrol inspection areas, and determining the patrol inspection priorities corresponding to the multiple patrol inspection areas based on historical data; Determining multiple subtask packages corresponding to the multiple patrol inspection areas, and determining the target subtask package corresponding to the drone according to the basic information of each drone; the basic information of the drone includes location information, remaining power and sensor information; Predicting the movement trajectory of the intelligent mobile device based on the intelligent mobile device information, and determining the patrol inspection path corresponding to the drone in combination with the preset 3D point cloud map corresponding to the high-risk enclosed scenario; Constructing a power task coupling model to predict the executable task duration of the drone based on the patrol inspection path, and after the drone executes the executable task duration, continuing the patrol inspection through a replaceable drone.

[0006] In one implementation manner of this application, before dividing the high-risk enclosed scenario into multiple patrol inspection areas, the method further includes: Integrate sensors onto the drone based on the standardized interfaces reserved on the drone, and collect scene data in high-risk enclosed scenarios through the sensors; the sensors include lidar, visual sensors, voiceprint sensors, and vibration sensors; Obtain the preset layout in the high-risk enclosed scenario, and generate a 3D point cloud map corresponding to the high-risk enclosed scenario based on the scene data and the preset layout; Compress the incremental data of the 3D point cloud map through edge computing to update the scene map in real time, and synchronize the updated scene map to other drones.

[0007] In an implementation manner of the present application, predict the movement trajectory of the intelligent mobile device based on the intelligent mobile device information, and combine the preset 3D point cloud map corresponding to the high-risk enclosed scenario to determine the inspection path corresponding to the drone, specifically including: Establish wireless communication between the drone and the intelligent mobile device in the corresponding inspection area to obtain the position information and movement plan corresponding to the intelligent mobile device; Predict the movement trajectory of the intelligent mobile device according to the position information and the movement plan, and determine the target point corresponding to the intelligent mobile device at each timestamp according to the movement trajectory; Determine the device priority of each intelligent mobile device according to the device type of the intelligent mobile device, and determine the inspection path corresponding to the drone based on the device priority.

[0008] In an implementation manner of the present application, determine the inspection path corresponding to the drone based on the device priority, specifically including: Determine whether the device priority of the intelligent mobile device is greater than that of the drone in the corresponding inspection area; If not, determine the same passing points of the drone and the intelligent mobile device to adjust the same passing points in the movement trajectory of the intelligent mobile device; If so, adjust the preset path of the drone to generate the target inspection path corresponding to the drone based on the remaining power of the drone and in combination with the 3D point cloud map corresponding to the high-risk enclosed scenario.

[0009] In an implementation manner of the present application, after constructing a power task coupling model to predict the executable task duration of the drone based on the inspection path, the method further includes: Determine the return charging timestamp and the corresponding return path of the drone based on the executable task duration of the drone; Mark the uncompleted inspection areas of the drone in the corresponding inspection area through visual heat map overlay; Generate the power decay curve of the drone during mission execution according to the mission execution situation of the drone, and calculate the remaining inspection task volume corresponding to the uncompleted inspection area.

[0010] In an implementation manner of the present application, after the drone executes the executable mission duration, continue the inspection through a replaceable drone, specifically including: Determine the replaceable drone corresponding to the inspection area according to the remaining inspection task volume, and continue to complete the remaining inspection task volume of the inspection area through the replaceable drone; Predict the charging demand of the drone according to the power decay curve and the task queue corresponding to the remaining inspection task volume, and trigger the drone charging pile to perform charging preheating based on the charging demand.

[0011] In an implementation manner of the present application, determine multiple sub-task packages corresponding to the multiple inspection areas, and determine the target sub-task package corresponding to the drone according to the basic information of each drone, specifically including: For each inspection area, split the inspection task of the inspection area into corresponding multiple sub-task packages, and determine the position information, remaining power, and sensor information of each drone; each sub-task package includes a task difficulty coefficient, an energy consumption weight, and the required sensor type, and the sensor information includes the sensor detection range and monitoring ability; Calculate the task benefit value of each drone respectively according to the position information, remaining power, and sensor information of the drone, and determine the drone corresponding to each sub-task package based on the task benefit value; the task benefit value is determined according to the task priority, energy consumption weight, and distance cost.

[0012] In an implementation manner of the present application, deploy a drone cluster and determine the corresponding drone type to perform dynamic teaming according to the drone type; the drone type includes long-endurance drones and high-payload drones; Through an encrypted communication protocol, share information with nearby scene drones within a preset range, so as to send a reinforcement request to the nearby scene drones when there is no replaceable drone in the high-risk closed scene; the shared information includes a 3D point cloud map and a mission log.

[0013] On the other hand, an embodiment of the present application also provides a drone inspection device for a high-risk closed scene, and the device includes: At least one processor; And a memory communicatively connected to the at least one processor; Among them, the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned drone inspection method for high-risk closed scenes.

[0014] On the other hand, an embodiment of the present application further provides a non-volatile computer storage medium storing computer executable instructions, which, when executed, implements a drone inspection method for high-risk closed scenes as described above.

[0015] The embodiments of the present application provide a drone inspection method, device, and medium for high-risk closed scenes, which at least have the following beneficial effects: By establishing a priority model based on historical leakage records, equipment aging data and environmental parameters, the inspection frequency of high-risk areas has been increased, significantly reducing the probability of sudden accidents. In addition, drones can quickly adjust inspection priorities in emergency situations, improving the dynamic monitoring coverage of hazardous sources. Combining the remaining power of drones with sensor types reduces the energy consumption of a single mission and improves equipment utilization. Through the task package conflict resolution algorithm, multiple machines are avoided from working in clusters, narrowing the standard deviation of regional task completion time. Dynamic path adjustment based on digital twin technology reduces the risk of collision between drones and equipment such as AGV transport vehicles, thereby improving the success rate of passage. By predicting the movement trajectory of personnel, real-time tracking and shooting of high-risk operations is achieved, improving the accuracy of identifying violations. Combined with the deployment of replaceable drones, the continuous operation time is extended. When a sudden leak occurs, the nearest drone can be dispatched based on the remaining power model, which improves the response speed compared to traditional solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 A flowchart of a drone inspection method for high-risk closed scenes provided in an embodiment of the present application; Figure 2 A schematic diagram of the internal structure of a drone inspection device for high-risk closed scenes provided in an embodiment of the present application. DETAILED DESCRIPTION

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments of this application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the scope of protection of this application.

[0018] The following will, with reference to the drawings, elaborate on the technical solutions provided by each embodiment of this application.

[0019] Figure 1 It is a schematic flowchart of a drone inspection method for high-risk enclosed scenarios provided by an embodiment of this application.

[0020] The implementation of the analysis method involved in the embodiments of this application can be a terminal device or a server, and this application does not impose special restrictions on this. For the convenience of understanding and description, the following embodiments will be described in detail using a server as an example.

[0021] It should be noted that this server can be a single device or a system composed of multiple devices, that is, a distributed server, and this application does not make specific limitations on this.

[0022] As Figure 1 shown, a drone inspection method for high-risk enclosed scenarios provided by an embodiment of this application includes: 101. Divide the high-risk enclosed scenario into multiple inspection areas, and determine the inspection priorities corresponding to the multiple inspection areas based on historical data.

[0023] Specifically, in an embodiment of this application, before dividing the high-risk enclosed scenario into multiple inspection areas, the method further includes: Integrate sensors onto the drone based on the standardized interfaces reserved on the drone, and collect scenario data in the high-risk enclosed scenario through the sensors; the sensors include lidar, vision sensors, voiceprint sensors, and vibration sensors; Obtain the preset layout in the high-risk enclosed scenario to generate a 3D point cloud map corresponding to the high-risk enclosed scenario according to the scenario data and the preset layout; Compress the incremental data of the 3D point cloud map through edge computing to update the scenario map in real time and synchronize the updated scenario map to other drones.

[0024] In one embodiment, a large chemical plant's underground storage tank area stores highly toxic chemicals such as acrylonitrile, presenting risks of toxic gas volatilization and pipeline corrosion and leakage. 30m³ storage tanks are densely arranged, and the inspection passage width is only 1.2m. The GPS signal is shielded in the underground environment, and the 4G network coverage rate is less than 30%, presenting environmental hazards, spatial complexity, and communication challenges.

[0025] Four groups of M12 aviation interfaces are reserved on the belly of the drone, which are respectively mounted with a lidar Velodyne VLP-16, a vision sensor Sony IMX586 binocular camera, a voiceprint sensor Infineon IM69D130, and a vibration sensor ADI ADXL357. When a suspected leak is detected, ground personnel insert the sensor module into the drone interface, and the system automatically identifies the sensor type and loads the corresponding driver, supporting plug-and-play.

[0026] The drone flies in a snake-shaped path, and the lidar generates 200,000 point cloud data per second. The vision sensor synchronously captures images of the storage tank surface to identify abnormal features such as rust and droplets. The real-time point cloud data is registered with the preset CAD drawings, i.e., the storage tank layout and pipeline orientation. The voiceprint sensor captures abnormal noises in the pipeline, and frequency analysis shows an abnormal peak at 5kHz, suspected of a micro-leak. A 3D semantic map with multiple data layers is generated, including the basic structure layer, i.e., the storage tank / pipeline position, the risk marking layer, i.e., the corrosion area and leakage point, and the dynamic obstacle layer, i.e., the real-time position of the AGV.

[0027] The edge node performs octree encoding and compression on the point cloud data. When the network is unobstructed, it uploads the complete data packet through 5G. When the network is blocked, it only transmits the differential data block. After the leading drone discovers a newly added corrosion area at the bottom of Tank No. 3, the edge node extracts the point cloud features of this area and broadcasts update instructions through LoRa networking. Other drones receive the instructions and locally update the map to avoid the corrosion area.

[0028] In one embodiment, in the scenario of a liquefied natural gas storage tank area, a LiDAR lidar, an infrared thermal imager, and an ultrasonic leakage sensor are integrated through the standardized interface of the drone. The drone scans the outer wall of the storage tank in a grid path. The LiDAR constructs the point cloud of the tank surface in real time, and the infrared sensor synchronously detects abnormal temperature rises at the flange connections. The edge computing node uses a compression algorithm based on the octree structure to compress the original point cloud data into a smaller incremental packet, and realizes millisecond-level map updates through the 5G network. When an abnormal temperature rise of 0.3°C is detected in a certain storage tank, the system immediately synchronizes the updated map of the hot spot area to the adjacent 3 drones, triggering the leakage review process, and shortening the response time compared with the traditional manual inspection.

[0029] 102. Determine multiple subtask packages corresponding to multiple inspection areas, and determine the target subtask package corresponding to the drone according to the basic information of each drone.

[0030] It should be noted that the basic information of the drone in the embodiment of the present application includes location information, remaining battery power, and sensor information.

[0031] Specifically, in an embodiment of the present application, determining multiple subtask packages corresponding to multiple inspection areas and determining the target subtask package corresponding to the drone according to the basic information of each drone specifically includes: For each inspection area, split the inspection task of the inspection area into corresponding multiple subtask packages, and determine the location information, remaining battery power, and sensor information of each drone; each subtask package includes a task difficulty coefficient, an energy consumption weight, and the required sensor type, and the sensor information includes the sensor detection range and monitoring ability; According to the location information, remaining battery power, and sensor information of the drone, calculate the task benefit value of each drone respectively, and based on the task benefit value, determine the drone corresponding to each subtask package; the task benefit value is determined according to the task priority, energy consumption weight, and distance cost.

[0032] In an embodiment, the roadway width of a fully mechanized coal mining face in a deep underground coal mine is 2.4 m, the height is 1.8 m, and there is a risk of collapse in local areas (roof-to-floor convergence > 50 mm / d). The gas concentration gradient distribution is 0.2% - 2%, and there are local high-temperature points. It is necessary to perform multi-modal detection on hydraulic supports, shearer drums, and gas sensors, covering an area > 1,200 m.

[0033] Split the 1,200 m roadway into 4 subtask packages according to the equipment distribution. The sensors include an infrared thermal imager, a high-definition video pan-tilt, a laser methane detector, and a millimeter-wave radar. The task benefit value = task priority / (energy consumption weight × distance cost), so as to determine the drone corresponding to each subtask package based on the task benefit value.

[0034] In an embodiment, by deploying a heterogeneous drone cluster, distinguishing long-endurance reconnaissance drones and high-payload disposal drones, dynamically forming teams as needed, sharing the computing power and data of the drone clusters in multiple high-risk scenarios, and realizing regional-level task reinforcement.

[0035] The drone clusters in adjacent scenarios share the environmental map and task logs through an encrypted communication protocol, and the reinforcement drones quickly take over high-priority tasks based on the shared data, including leakage adsorption and plugging agent injection.

[0036] 103. Predict the movement trajectory of the intelligent mobile device based on the intelligent mobile device information, and combine the preset 3D point cloud map corresponding to the high-risk closed scenario to determine the inspection path corresponding to the drone.

[0037] Specifically, in an embodiment of the present application, the movement trajectory of the intelligent mobile device is predicted based on the intelligent mobile device information, and the corresponding inspection path of the drone is determined in combination with the preset 3D point cloud map corresponding to the high-risk enclosed scenario, specifically including: Establish wireless communication between the drone and the intelligent mobile device in the corresponding inspection area to obtain the position information and motion plan corresponding to the intelligent mobile device; According to the position information and motion plan, predict the movement trajectory of the intelligent mobile device, and determine the target point corresponding to the intelligent mobile device at each timestamp according to the movement trajectory; According to the device type of the intelligent mobile device, determine the device priority of each intelligent mobile device, and determine the corresponding inspection path of the drone based on the device priority.

[0038] In an embodiment, in the e-commerce logistics warehousing scenario, the drone establishes communication with the AGV handling robot and the warehousing robotic arm through Bluetooth. The AGV sends real-time motion planning data including the target shelf coordinates and the estimated arrival time. The drone uses the extended Kalman filter algorithm to predict its future 15-second trajectory and generates a target point sequence including 20 timestamps. The system sets the priority according to the device type: the priority of the robotic arm is 1, which involves cargo grabbing operations, the priority of the AGV is 2, and the priority of the conveyor belt is 3. The drone path planning module uses the dynamic A* algorithm to set an electronic fence within a radius of 0.8 meters of the robotic arm's working radius. When it is predicted that there is a trajectory conflict between the AGV and the robotic arm, the inspection path is automatically adjusted to the standby channel, reducing the interference rate of the warehousing operation.

[0039] In an embodiment, based on the device priority, determining the corresponding inspection path of the drone specifically includes: Determine whether the device priority of the intelligent mobile device is greater than that of the drone in the corresponding inspection area; If not, determine the same passing points of the drone and the intelligent mobile device to adjust the same passing points in the movement trajectory of the intelligent mobile device; If so, adjust the preset path of the drone to generate the corresponding target inspection path of the drone based on the remaining battery power of the drone and in combination with the 3D point cloud map corresponding to the high-risk enclosed scenario.

[0040] In an embodiment, in the nuclear reactor maintenance workshop, the drone performs infrared temperature measurement of the steam pipeline (priority 2) and cooperates with the welding robot (priority 1) and the inspection robotic arm (priority 3). When the welding robot starts the operation of thick-weld seams, it is determined that the priority is welding robot > drone > robotic arm.

[0041] Detect the working area of the robot based on the 3D point cloud map. A spherical area with a radius of 3m in the working area is a no-fly zone. The RRT* algorithm is used to generate a detour path, and the inspection order is optimized by combining the remaining battery power. For example, the original path is Area A → Area B → Area C → Area D, and the adjusted path is Area A → Area D → Area C → Area B, avoiding the welding operation area in Area B. And a visual sensor is enabled for auxiliary positioning to compensate for the path increment caused by the detour.

[0042] 104. Build a power-task coupling model to predict the executable task duration of the UAV based on the inspection path. After the UAV executes the executable task duration, continue the inspection through a replaceable UAV.

[0043] Specifically, in an embodiment of the present application, after building a power-task coupling model to predict the executable task duration of the UAV based on the inspection path, the method further includes: Determine the return charging timestamp and the corresponding return path of the UAV based on the executable task duration of the UAV; Mark the uncompleted inspection areas of the UAV in the corresponding inspection area through visual heat map overlay; Generate a power decay curve of the UAV during the task execution according to the task execution situation of the UAV, and calculate the remaining inspection task volume corresponding to the uncompleted inspection area.

[0044] In one embodiment, there are 12 normal temperature and pressure storage tanks in the storage tank area of a large chemical industrial park, storing flammable and explosive liquids. The distance between the storage tanks is 15 - 20 meters. Some areas are blocked by pipeline galleries, forming a complex three-dimensional structure, with combustible gas leakage and high-temperature areas. It is necessary to detect millimeter-level cracks in the welds, flange interfaces and safety valves of the storage tanks every day, and the coverage area > 20,000 ㎡.

[0045] The UAV is equipped with an infrared thermal imager and a laser methane detector, with a cruising speed of 4m / s and a total battery capacity of 6000mAh. The remaining executable task duration = (current power - safety threshold power) / (sensor power consumption + flight power consumption). When the remaining power ≤ 20%, with about 12 minutes of endurance and the uncompleted inspection area > 10%, start the return flight program.

[0046] Combined with the gradient information in the 3D point cloud map, an improved A* algorithm is used to avoid steam pipelines and leakage areas. When the UAV returns from the west side of the No. 5 storage tank, it detects that the VOC concentration exceeds the standard 20 meters in the southeast direction. The system automatically adjusts the path to the safe passage on the north side, increasing the return flight distance by 15m but reducing the collision risk.

[0047] When visualizing the uncompleted areas with a heat map, red indicates uninspected areas with a priority level ≥ 8, orange indicates uninspected areas with a priority level of 5 - 7, and green indicates areas that have been inspected and completed. The heat map is fused with the 3D point cloud map to mark leakage points, high-temperature areas, and obstacles.

[0048] When UAV A returned due to insufficient power while inspecting the east side of Tank 3, the heat map showed that there were two flange connections in this area (area ≈ 300㎡) that had not been detected. The system immediately dispatched UAV B equipped with the same type of sensor to preferentially cover the red-marked area.

[0049] When calculating the power decay curve and task volume of the UAV, the infrared thermal imager shooting interval and laser methane detection frequency are recorded. Through historical data fitting, the relationship between the power consumption rate, sensor usage frequency, and flight altitude is obtained. For example, when UAV B took over the task, 200 infrared images needed to be taken and 500 groups of methane data needed to be collected for the remaining uninspected area on the east side of Tank 3, with an estimated time consumption of 18 minutes.

[0050] According to the power decay curve, the charging pile is triggered for preheating, and the ground personnel are notified to prepare for battery replacement. If UAV B has insufficient power and there is no replaceable UAV, the system automatically sends a reinforcement request to the fire UAV cluster in the adjacent scene, and preferentially dispatches a standby machine equipped with a high-definition video transmission module. By simulating the impact of different sensor combinations on power consumption through the digital twin platform, the task allocation strategy is optimized.

[0051] In one embodiment, after the UAV has executed the executable task duration, the inspection is continued by a replaceable UAV, which specifically includes: Determine the replaceable UAV corresponding to the inspection area according to the remaining inspection task volume, and continue to complete the remaining inspection task volume of the inspection area by the replaceable UAV; Based on the power decay curve and the task queue corresponding to the remaining inspection task volume, predict the charging demand of the UAV, and trigger the UAV charging pile for charging preheating based on the charging demand.

[0052] In one embodiment, in a reactor building of a certain third-generation nuclear power plant, the red and orange areas are distributed in layers, and dynamic path planning is required to avoid high-radiation areas. It includes steam generator compartments, ring crane tracks, and coolant pipes. It is necessary to detect the surface cracks of the reactor pressure vessel, the weld status of the steam pipes, and the radiation dose distribution in real time, with a coverage area > 5,000㎡.

[0053] The uninspected areas are sorted according to the radiation dose and equipment criticality at the risk level to generate a 3D heat map (red / orange / yellow / green). When the main drone's battery level ≤ 15% or a sudden change in radiation dose is detected, i.e., > 20% of the preset threshold, the system automatically calculates the remaining task volume. For example, the area of the uninspected area is 400㎡, including 2 red areas, and 120 high-definition videos + 200 sets of radiation data need to be captured, with an estimated time consumption of 22 minutes.

[0054] The backup drone receives the real-time 3D map, the coordinates of the uninspected area, and sensor parameters such as the infrared temperature measurement threshold through the 5G network. Based on the radiation field simulation of the digital twin platform, it plans to avoid newly emerging high-radiation areas, such as a sudden leakage point of a certain steam pipeline.

[0055] Combined with the current battery level, the task queue duration, sensor power consumption, and flight mode, predict the timestamp when the battery runs out. The charging duration required = estimated time consumption - available duration of the remaining battery = 30 - 28 = 2 minutes, triggering the "emergency charging" flag.

[0056] The above is the method embodiment proposed in this application. Based on the same inventive concept, the embodiment of this application also provides a drone inspection device for high-risk enclosed scenarios, and its structure is as Figure 2 shown.

[0057] Figure 2 This is the internal structure schematic diagram of a drone inspection device for high-risk enclosed scenarios provided by the embodiment of this application. As Figure 2 shown, the device includes: At least one processor; And a memory communicatively connected to at least one processor; Wherein, the memory stores instructions executable by at least one processor, and the instructions are executed by at least one processor to enable at least one processor to: Divide the high-risk enclosed scenario into multiple inspection areas, and determine the inspection priorities corresponding to the multiple inspection areas based on historical data; Determine multiple subtask packages corresponding to the multiple inspection areas, and determine the target subtask package corresponding to the drone according to the basic information of each drone; the basic information of the drone includes location information, remaining battery level, and sensor information; Predict the movement trajectory of the intelligent mobile device based on the intelligent mobile device information, and combine the preset 3D point cloud map corresponding to the high-risk enclosed scenario to determine the inspection path corresponding to the drone; Construct a power-task coupling model to predict the executable task duration of the drone based on the inspection path, and after the drone executes the executable task duration, continue the inspection through a replaceable drone.

[0058] The embodiments of the present application also provide a non-volatile computer storage medium storing computer-executable instructions, which can perform the following operations when executed: Divide a high-risk closed scenario into multiple inspection areas, and determine the inspection priorities corresponding to the multiple inspection areas based on historical data; Determine multiple subtask packages corresponding to the multiple inspection areas, and determine the target subtask package corresponding to the drone according to the basic information of each drone; the basic information of the drone includes location information, remaining battery power, and sensor information; Predict the movement trajectory of the intelligent mobile device based on the intelligent mobile device information, and determine the inspection path corresponding to the drone in combination with the preset 3D point cloud map corresponding to the high-risk closed scenario; Construct a power-task coupling model to predict the executable task duration of the drone based on the inspection path, and after the drone executes the executable task duration, continue the inspection through a replaceable drone.

[0059] The embodiments in the present application are all described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the differences between each embodiment and other embodiments are emphasized. In particular, for the device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiments.

[0060] The devices and media provided by the embodiments of the present application correspond one by one to the methods. Therefore, the devices and media also have beneficial technical effects similar to those of the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be elaborated here.

[0061] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0062] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, as well as the combination of flows and / or blocks in the flowchart and / or block diagram. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0063] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0064] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0065] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0066] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. The memory is an example of computer-readable media.

[0067] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0068] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0069] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A drone inspection method for high-risk closed scenes, characterized in that: The method comprises: Divide the high-risk closed scene into multiple inspection areas, and determine the inspection priorities corresponding to the multiple inspection areas based on historical data; Determine multiple subtask packages corresponding to the multiple inspection areas, and determine the target subtask package corresponding to the drone according to the basic information of each drone; the basic information of the drone includes location information, remaining power and sensor information; Predicting the movement trajectory of the smart mobile device based on the smart mobile device information, and determining the inspection path corresponding to the drone in combination with the preset 3D point cloud map corresponding to the high-risk closed scene; A power-task coupling model is constructed to predict the executable task duration of the UAV based on the inspection path, and after the UAV executes the executable task duration, the inspection is continued by a replaceable UAV.

2. The method for drone inspection of high-risk closed scenes according to claim 1 is characterized in that: Before dividing the high-risk closed scene into a plurality of inspection areas, the method further includes: Based on the standardized interface reserved on the drone, the sensor is integrated into the drone, and the scene data in the high-risk closed scene is collected through the sensor; the sensor includes a laser radar, a visual sensor, a voiceprint sensor and a vibration sensor; Acquire a preset layout in the high-risk enclosed scene to generate a 3D point cloud map corresponding to the high-risk enclosed scene according to the scene data and the preset layout; The incremental data of the 3D point cloud map is compressed through edge computing to update the scene map in real time, and the updated scene map is synchronized to other drones.

3. The method for drone inspection of high-risk closed scenes according to claim 1 is characterized in that: Predicting the movement trajectory of the smart mobile device based on the smart mobile device information, and combining the preset 3D point cloud map corresponding to the high-risk closed scene to determine the inspection path corresponding to the drone, specifically including: Establishing wireless communication between the drone and the smart mobile device in the corresponding inspection area to obtain the location information and motion plan corresponding to the smart mobile device; Predicting the movement trajectory of the intelligent mobile device according to the location information and the motion plan, and determining the target point position of the intelligent mobile device corresponding to each timestamp according to the movement trajectory; According to the device type of the smart mobile device, the device priority of each smart mobile device is determined, and based on the device priority, the inspection path corresponding to the drone is determined.

4. The method for unmanned aerial vehicle inspection for high-risk closed scenes according to claim 3 is characterized in that: Based on the device priority, the inspection path corresponding to the drone is determined, specifically including: Determine whether the device priority of the smart mobile device is greater than the drone in the corresponding inspection area; If not, determining the common waypoints of the drone and the smart mobile device to adjust the common waypoints in the moving trajectory of the smart mobile device; If so, the preset path of the drone is adjusted to generate a target inspection path corresponding to the drone based on the remaining power of the drone and in combination with the 3D point cloud map corresponding to the high-risk closed scene.

5. The method for drone inspection of high-risk closed scenes according to claim 1 is characterized in that: After constructing a power-task coupling model to predict the duration of the mission that can be performed by the drone based on the inspection path, the method further includes: Determine the return charging timestamp and the corresponding return path of the drone based on the executable mission duration of the drone; By overlaying a visualized heat map, marking the unfinished inspection area of ​​the drone in the corresponding inspection area; According to the mission execution status of the drone, a power attenuation curve of the drone during the mission execution is generated, and the remaining inspection task amount corresponding to the unfinished inspection area is calculated.

6. The method for drone inspection of high-risk closed scenes according to claim 5 is characterized in that: After the drone performs the executable task for a certain period of time, the inspection is continued by using a replaceable drone, specifically including: Determine a replaceable drone corresponding to the inspection area according to the remaining inspection task volume, and continue to complete the remaining inspection task volume of the inspection area through the replaceable drone; According to the power attenuation curve and the task queue corresponding to the remaining inspection task amount, the charging demand of the drone is predicted, and based on the charging demand, the drone charging pile is triggered to perform charging preheating.

7. The method for drone inspection of high-risk closed scenes according to claim 1 is characterized in that: Determining multiple subtask packages corresponding to the multiple inspection areas, and determining a target subtask package corresponding to the drone according to basic information of each drone, specifically including: For each inspection area, the inspection task of the inspection area is divided into corresponding multiple subtask packages, and the location information, remaining power and sensor information of each drone are determined; each subtask package includes the task difficulty coefficient, energy consumption weight and required sensor type, and the sensor information includes sensor detection range and monitoring capability; According to the location information, remaining power and sensor information of the drone, the mission benefit value of each drone is calculated respectively, and based on the mission benefit value, the drone corresponding to each subtask package is determined; the mission benefit value is determined according to the task priority, energy consumption weight and distance cost.

8. The method for drone inspection of high-risk closed scenes according to claim 1 is characterized in that: The method further comprises: Deploy a drone cluster and determine the corresponding drone type to dynamically form a team based on the drone type; the drone type includes a long-endurance drone and a high-payload drone; Through an encrypted communication protocol, information is shared with nearby scene drones within a preset range, so that when there are no replacement drones in the high-risk closed scene, a reinforcement request is sent to the nearby scene drones; the shared information includes 3D point cloud maps and mission logs.

9. A drone inspection device for high-risk closed scenes, characterized in that: The device comprises: at least one processor; and, a memory communicatively coupled to the at least one processor; In which, the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a drone inspection method for high-risk closed scenes as described in any one of claims 1-8.

10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: When the computer executable instructions are executed, a drone inspection method for high-risk closed scenes as described in any one of claims 1 to 8 is implemented.

Images