Unmanned aerial vehicle river and lake patrol path planning method and system based on Beidou navigation

Abstract

The invention provides an unmanned aerial vehicle river and lake patrol path planning method and system based on Beidou navigation, and relates to the field of unmanned aerial vehicle patrol. According to the method, real-time position information of the unmanned aerial vehicle is obtained, and whether the unmanned aerial vehicle meets data acquisition requirements is judged based on preset key monitoring point information. And if not, acquiring environment sensing data, and constructing a safe flight area of the unmanned aerial vehicle. And calculating temporary monitoring points meeting data acquisition requirements, and constructing a river and lake patrol flight path between the unmanned aerial vehicle and the temporary monitoring points. The method effectively solves the problems of low patrol efficiency and even possible missing of important monitoring points caused by the fact that the patrol route cannot be flexibly adjusted in the traditional unmanned aerial vehicle patrol path planning. By dynamically adjusting the path in real time, the defects of providing high-precision real-time positioning and optimizing the path in the prior art are overcome, the flexibility and efficiency of patrol and the accuracy of data acquisition are remarkably improved, task interruption is avoided, and successful data acquisition of a key area is ensured.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) patrol technology, and more specifically, to a method and system for planning UAV river and lake patrol routes based on BeiDou navigation. Background Technology

[0002] In the daily monitoring and maintenance of river and lake environments, drone patrols have become an important tool. However, traditional drone patrol methods often rely on pre-set flight routes, which prove inadequate when facing complex and ever-changing aquatic environments and unexpected situations. For example, when the water morphology changes or an emergency occurs, drones need to be able to flexibly adjust their patrol routes to promptly cover areas requiring close monitoring. However, existing technologies are insufficient in providing high-precision real-time positioning and optimized routes, often resulting in low patrol efficiency and potentially missing important monitoring points.

[0003] Specifically, in the actual operation of river and lake water environment monitoring and maintenance, drone patrol systems typically operate according to preset flight paths. These paths are planned by comprehensively considering the geographical features of the water area, key monitoring areas, and known obstacle information. However, with continuous infrastructure development and environmental changes, drone patrols face the challenge of path planning failing to adapt in real time to complex water features and sudden environmental changes. For example, when encountering sudden water pollution or extreme weather, drones need to dynamically adjust their patrol paths to cover key areas, but existing technologies have limitations in high-precision real-time positioning and path optimization, potentially leading to low patrol efficiency or the omission of key monitoring points.

[0004] When drones conduct river and lake patrols, they face multiple interconnected navigation information failures in complex aquatic environments. These include untimely map data updates, unreliable positioning information due to severe interference with BeiDou navigation signals, and rapid accumulation and drift of inertial navigation system errors. Existing path planning methods struggle to dynamically and adaptively generate a new route that effectively avoids unidentified obstacles while meeting the patrol mission objectives within a very short time, based on limited and uncertain perception information. This is crucial to prevent mission interruption and ensure successful data collection in critical areas. Therefore, existing technologies urgently need improvement to address these issues. Summary of the Invention

[0005] This application discloses a method and system for planning the route of UAV river and lake patrol based on Beidou navigation. It aims to solve the problems of traditional UAV patrol route planning, which cannot flexibly adjust the patrol route when facing complex and changeable water environment and emergencies, resulting in low patrol efficiency and even the possibility of missing important monitoring points, as well as the difficulty in dynamically and adaptively generating new routes when navigation information fails.

[0006] The first objective of this invention is to propose a method for unmanned aerial vehicle (UAV) river and lake patrol path planning based on BeiDou navigation, comprising: Obtain the real-time location information of the drone, and based on the preset key monitoring point information and the drone's real-time location information, determine whether the drone meets the data collection requirements of the key monitoring points; If it is determined that the drone does not meet the data collection requirements of the key monitoring points, environmental perception data around the drone is collected through Beidou navigation, and based on the environmental perception data, a safe flight area for the drone is constructed within a spatial range centered on the drone. Based on key monitoring point information and the safe flight area of ​​UAVs, temporary monitoring points are calculated to ensure that the real-time location of the UAV meets the data collection requirements of the key monitoring points, and a river and lake patrol flight path is constructed between the UAV and the temporary monitoring points.

[0007] This solution enables drones to dynamically determine data collection needs during river and lake patrols based on real-time location and key monitoring point information. When the requirements are not met, a safe flight zone can be constructed using BeiDou navigation and environmental perception data. This allows for the planning of temporary monitoring points and flight paths that meet the data collection requirements, effectively solving the limitations of traditional patrol path planning and improving the flexibility and efficiency of patrols.

[0008] Preferably, the real-time location information of the UAV is acquired, and based on the preset key monitoring point information and the real-time location information of the UAV, it is determined whether the UAV meets the data collection requirements of the key monitoring points, including: Obtain real-time location information of the drone, either in real-time or at regular intervals; Obtain information on preset key monitoring points, including the geographic coordinates, optimal observation height range, and optimal observation angle range of all key monitoring points in the patrol task. Based on the geographic coordinates of all key monitoring points and the real-time location of the UAV, the relative distance and relative angle between the current location of the UAV and each preset key monitoring point are calculated. The relative distance between the current location of the drone and each preset key monitoring point is compared with the optimal observation altitude range, and the relative angle between the current location of the drone and each preset key monitoring point is compared with the optimal observation angle range. If the relative distance between the current location of the drone and one or more preset key monitoring points exceeds the optimal observation height range, or the relative angle between the current location of the drone and one or more preset key monitoring points exceeds the optimal observation angle range, then the drone is determined not to meet the data collection requirements of the key monitoring points. The system identifies key monitoring points where the drone does not meet data collection requirements, and also identifies the deviation between the drone's current location and these key monitoring points.

[0009] This solution enables accurate determination of whether a UAV meets the data collection requirements of key monitoring points. By comparing the UAV's position with the optimal observation parameters of key monitoring points in detail and marking deviation information, it provides accurate input for subsequent path planning, thereby improving the accuracy and effectiveness of data collection.

[0010] Preferably, environmental perception data around the UAV is collected via BeiDou navigation, and based on this data, a safe flight zone for the UAV is constructed within a spatial range centered on the UAV, including: By collecting environmental perception data around the drone using BeiDou navigation, obstacle information can be obtained within the spatial range centered on the drone. Based on environmental perception data around the drone and obstacle information within the space centered on the drone, a safe flight zone for drones is constructed within the space centered on the drone.

[0011] This solution utilizes the BeiDou Navigation Satellite System to acquire precise environmental perception data and identify obstacles based on this data, thereby constructing a safe flight zone centered on the UAV. This effectively ensures the flight safety of the UAV in complex environments and avoids the risk of collisions.

[0012] Preferably, based on key monitoring point information and the safe flight area of ​​the UAV, temporary monitoring points are calculated to ensure that the real-time position of the UAV meets the data collection requirements of the key monitoring points. A river and lake patrol flight path is then constructed between the UAV and the temporary monitoring points, including: Starting from the drone's current location, sampling points are randomly set within the drone's safe flight area, and the nearest key monitoring point is found among the key monitoring points that is closest to the sampling point; Temporary monitoring points are generated by extending a preset step length from the nearest key monitoring point toward the sampling point. When using temporary monitoring points as key monitoring points, check whether the drones meet the data collection requirements of the key monitoring points; If the conditions are met, establish a river and lake patrol flight path from the current location of the drone to the temporary monitoring point; if not, continue to set new sampling points and execute the aforementioned steps until a temporary monitoring point that meets the data collection requirements of the key monitoring point is obtained.

[0013] This solution enables the dynamic and adaptive planning of temporary monitoring points and flight paths that meet data collection requirements. Through iterative searching and verification within a safe flight area, it ensures that the UAV can efficiently reach and collect data from key monitoring points, significantly improving the flexibility and success rate of patrol missions.

[0014] Preferably, after the step of constructing the flight path between the UAV and the temporary monitoring point, the method further includes a step of optimizing the flight path by adaptively adjusting the onboard sensor parameters of the UAV, including: Focus adjustment: If the distance between the UAV and the temporary monitoring point in the flight path is greater than the first preset distance, the focus parameter of the UAV's onboard image acquisition sensor will be adaptively adjusted. Exposure adjustment: If the light intensity changes during the flight of the drone along the flight path, the exposure time and sensitivity of the drone's onboard image acquisition sensor will be adaptively adjusted. Frequency adjustment: If the UAV's flight speed increases or the sampling frequency increases, the scanning frequency of the multispectral sensor will be adaptively adjusted. Exposure time adjustment for different spectral bands: Adaptively adjust the exposure time for each spectral band based on the differences in sensitivity to light intensity among different spectral bands; Image quality adjustment: The image quality acquired by the image acquisition sensor is evaluated. If the image sharpness or the signal-to-noise ratio of the spectral data does not meet the preset requirements, the shooting parameters of the image acquisition sensor are adaptively fine-tuned.

[0015] This solution adaptively adjusts the parameters of airborne sensors based on flight path and environmental changes, ensuring that the quality of image and spectral data acquisition remains optimal under different flight conditions and lighting environments, thereby significantly improving the reliability and effectiveness of patrol data.

[0016] Preferably, the flight path is optimized by adaptively adjusting the parameters of the UAV's onboard sensors, and the image acquisition sensor's shooting parameters are also adaptively adjusted to obtain the optimal images of key monitoring points, including: Illumination difference detection is performed on key monitoring point images acquired by image acquisition sensors, and abnormal illumination regions in the key monitoring point images are identified based on the brightness difference detection results; among them, abnormal illumination regions include overexposed regions and underexposed regions; By fine-tuning the drone's flight attitude parameters, the shooting angle and / or shooting distance of the image acquisition sensor can be adaptively adjusted. After each fine-tuning of the flight attitude parameters, the UAV continuously acquires multiple images of key monitoring points based on the image acquisition sensor. All the images of key monitoring points form a key monitoring point image sequence. The key monitoring point image sequence contains multiple images of the same area under different lighting angles. The key monitoring point image sequence is aligned and analyzed to obtain the optimal key monitoring point image.

[0017] This solution effectively overcomes the limitations of a single image under complex lighting conditions by combining refined illumination difference detection and flight attitude fine-tuning with the acquisition and analysis of multiple image sequences. This results in high-quality images of the optimal key monitoring points, greatly improving the accuracy and reliability of data acquisition.

[0018] Preferably, the key monitoring point image sequence is aligned and analyzed to obtain the optimal key monitoring point image, including: A feature-point-based image registration algorithm is used to align the image sequence of key monitoring points in order to compensate for the image displacement caused by fine-tuning the attitude of the UAV. A multi-exposure image fusion algorithm was used to analyze the aligned key monitoring point image sequence. One key monitoring point image was selected from the aligned key monitoring point image sequence as the initial optimal key monitoring point image. For overexposed areas in the initial optimal key monitoring point image, select the same areas with weak reflection and clear details in the key monitoring point images other than the initial optimal key monitoring point image in the aligned key monitoring point image sequence to replace them. For underexposed areas in the initial optimal key monitoring point image, select the same areas with high brightness and obvious texture in the key monitoring point images other than the initial optimal key monitoring point image in the key monitoring point image sequence to replace them. The key monitoring point image obtained after substitution is taken as the optimal key monitoring point image.

[0019] This solution effectively handles displacement and illumination differences in image sequences through image registration and multi-exposure fusion technology. By replacing overexposed and underexposed areas with areas that are clearer in detail and have more moderate brightness through region substitution, it generates a key monitoring point image with the best overall quality, significantly improving the usability and analytical value of image data.

[0020] Preferably, the adaptive adjustment of the onboard sensor parameters of the UAV further includes adaptive adjustment of the water quality sensor parameters, including: When the drone receives a water quality data collection command, it controls the water quality sensor to simultaneously collect water quality parameters in the water body at key monitoring points. The types of water quality parameters include pH value, conductivity, turbidity, and dissolved oxygen content. Real-time monitoring of each type of water quality parameter, statistical analysis of the instantaneous change rate of each type of water quality parameter, and determination of whether there are abnormal fluctuations in the instantaneous change rate of each type of water quality parameter. If the instantaneous change rate of two or fewer types of water quality parameters fluctuates abnormally at the same time, it is determined that the water quality at the key monitoring point has not changed substantially. The water quality parameters are then calibrated and filtered, and the calibrated and filtered water quality parameters are transmitted to the host system through the communication system of the drone. If the instantaneous change rate of two or more types of water quality parameters fluctuates abnormally at the same time, and it is determined that the water quality at the key monitoring point has undergone substantial changes, then the raw data of the water quality parameters will be transmitted to the host system through the communication system of the drone.

[0021] This application enables intelligent monitoring and judgment of water quality parameters. By analyzing the instantaneous change rate of various water quality parameters, it distinguishes between substantial and non-substantial changes in water quality and performs corresponding data processing and transmission, thereby improving the accuracy of water quality monitoring and the timeliness of decision-making.

[0022] Preferably, before the step of adaptively adjusting the image acquisition sensor's shooting parameters, the method further includes: The drone's onboard environmental perception sensors collect environmental visibility data of the drone's location and compare it with a preset visibility threshold. When the environmental visibility data collected by the environmental perception sensor is lower than the visibility threshold, the drone performs a rough location determination of the current location and combines it with the historical map information stored in the drone's onboard system to determine the prior scene structure that best matches the drone's current location. The geometric parameters of the prior scene structure are adjusted and updated by using sparse point cloud data collected by UAV-borne lidar and blurred visual images collected by image acquisition sensors under low visibility conditions. Based on the adjustment and update of the geometric parameters of the prior scene structure, an environmental topology graph of the prior scene structure is constructed; where nodes in the environmental topology graph represent the central region of key monitoring points, and edges in the environmental topology graph represent feasible paths between nodes.

[0023] This solution effectively addresses the challenges of drone patrols in low-visibility environments. By combining data from multiple sensors and historical map information, it constructs and updates an environmental topology map, ensuring that drones can still perform accurate positioning and path planning under adverse weather conditions, greatly improving the robustness and adaptability of patrol missions.

[0024] The second objective of this invention is to propose a UAV river and lake patrol route planning system based on BeiDou navigation, comprising: The judgment module is used to obtain the real-time location information of the UAV and, based on the preset key monitoring point information and the real-time location information of the UAV, determine whether the UAV meets the data collection requirements of the key monitoring points. The UAV safe flight area construction module is used to collect environmental perception data around the UAV through Beidou navigation when it is determined that the UAV does not meet the data collection requirements of key monitoring points, and to construct a safe flight area for the UAV within a spatial range centered on the UAV based on the environmental perception data. The flight path construction module is used to calculate temporary monitoring points that meet the data collection requirements of key monitoring points based on key monitoring point information and the safe flight area of ​​UAVs, and to construct the river and lake patrol flight path between the UAV and the temporary monitoring points.

[0025] This application provides a system-level solution through this technical solution. Through modular design, it realizes the automation and intelligence of UAV river and lake patrol path planning, which can efficiently determine data collection needs, construct safe flight areas and plan flight paths, thereby improving the performance and reliability of the entire patrol system.

[0026] Unlike existing technologies, this application provides a method for planning UAV river and lake patrol routes based on BeiDou navigation. It acquires the UAV's real-time location information and determines whether the UAV meets data collection requirements based on preset key monitoring point information. If not, it uses BeiDou navigation to collect environmental perception data and constructs a safe flight zone for the UAV. Based on this, it calculates temporary monitoring points that meet the data collection requirements and constructs a river and lake patrol flight path between the UAV and these temporary monitoring points. This method effectively solves the problem that traditional UAV patrol route planning cannot flexibly adjust patrol routes when facing complex and changing aquatic environments and emergencies, resulting in low patrol efficiency and even the potential omission of important monitoring points. By dynamically adjusting the path in real time, this application overcomes the shortcomings of existing technologies in providing high-precision real-time positioning and optimizing paths, significantly improving the flexibility, efficiency, and accuracy of patrol data collection, avoiding mission interruptions, and ensuring successful data collection in key areas.

[0027] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0028] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a method for planning the path of a UAV river and lake patrol based on BeiDou navigation, provided by the present invention.

[0029] Figure 2 This is a schematic diagram of the structure of a UAV river and lake patrol path planning system based on Beidou navigation provided by the present invention. Detailed Implementation

[0030] Embodiments of the present invention are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0031] In the daily monitoring and maintenance of river and lake environments, drone patrols have become an important tool. However, traditional drone patrol methods often rely on pre-set flight routes, which prove inadequate when facing complex and ever-changing aquatic environments and unexpected situations. For example, when the water morphology changes or an emergency occurs, drones need to be able to flexibly adjust their patrol routes to promptly cover areas requiring close monitoring. However, existing technologies are insufficient in providing high-precision real-time positioning and optimized routes, often resulting in low patrol efficiency and potentially missing important monitoring points.

[0032] like Figure 1 As shown, this embodiment of the invention provides a method for unmanned aerial vehicle (UAV) river and lake patrol path planning based on BeiDou navigation, including: S110: Obtain the real-time location information of the UAV, and determine whether the UAV meets the data collection requirements of the key monitoring points based on the preset key monitoring point information and the real-time location information of the UAV.

[0033] In embodiments of the present invention, the real-time location information of the UAV refers to the precise geographic coordinates of the UAV at a certain moment, which is usually provided by the UAV's onboard positioning system (such as the BeiDou Navigation Satellite System). The preset key monitoring point information refers to the geographical locations that are pre-set in river and lake patrol missions and require focused attention and data collection; this information may include geographic coordinates, optimal observation altitude, and angle, etc.

[0034] In practical implementation, this method first requires acquiring the real-time location information of the UAV. This can be achieved in several ways. For example, the UAV can be equipped with a BeiDou navigation receiver module, which can receive BeiDou satellite signals in real time and calculate the UAV's precise three-dimensional position coordinates (longitude, latitude, and altitude). This location information can be sent to the UAV control system at fixed time intervals (e.g., once per second). Another approach is for the UAV to integrate an inertial measurement unit (IMU) and a visual odometry system, fusing multi-sensor data to estimate its own position and periodically calibrating with BeiDou navigation data to improve positioning accuracy and robustness.

[0035] Based on pre-set key monitoring point information and the UAV's real-time location information, it is necessary to determine whether the UAV meets the data acquisition requirements of the key monitoring points. The pre-set key monitoring point information can be stored in advance in the UAV's onboard system or ground control station, including the geographic coordinates, optimal observation altitude range, and optimal observation angle range for each key monitoring point. The determination process involves comparing the UAV's current position with the geographic coordinates of each key monitoring point to calculate the relative distance and relative angle. For example, if the relative distance between the UAV and a certain key monitoring point exceeds the optimal observation altitude range for that monitoring point, or the relative angle exceeds the optimal observation angle range, it can be determined that the UAV does not meet the data acquisition requirements for that key monitoring point.

[0036] To ensure the quality of data collection for drone river and lake patrol missions, it is necessary to accurately determine whether the drone meets the data collection requirements of key monitoring points. Specifically, determining whether a drone meets the data collection requirements of key monitoring points includes the following steps: Obtain real-time location information of the drone, either in real-time or at regular intervals.

[0037] Obtaining real-time or periodically transmitted drone location information involves the drone periodically transmitting its current geographic coordinates (including longitude, latitude, and altitude) to the onboard control system or ground control station via its onboard positioning module, such as a BeiDou navigation module, a Global Positioning System (GPS) module, or other satellite navigation system module, or when triggered by specific events. This real-time or periodically transmitted data ensures continuous monitoring of the drone's position.

[0038] Obtain information on preset key monitoring points, including the geographic coordinates, optimal observation height range, and optimal observation angle range of all key monitoring points in the patrol task.

[0039] Acquiring pre-set key monitoring point information involves pre-configuring and storing data on all important monitoring points involved in the river and lake patrol mission within the UAV's onboard system or host system before the mission begins. This pre-set key monitoring point information not only includes the precise geographic coordinates of each key monitoring point but also further refines the optimal observation altitude and angle ranges required for effective data acquisition. The optimal observation altitude range can be understood as the ideal distance interval between the UAV and the key monitoring point in the vertical direction, ensuring that the sensors can capture clear and comprehensive images or data. The optimal observation angle range refers to the ideal observation angle interval between the UAV and the key monitoring point in the horizontal direction, avoiding issues such as obstruction, distortion, or uneven lighting, thereby ensuring data quality.

[0040] Based on the geographic coordinates of all key monitoring points and the real-time location of the UAV, the relative distance and relative angle between the current location of the UAV and each preset key monitoring point are calculated.

[0041] This method uses geometric calculations to compare the real-time 3D position of the UAV with the 3D geographic coordinates of each key monitoring point. The relative distance is the Euclidean distance between the UAV and the key monitoring point, while the relative angle can include the horizontal azimuth angle and the vertical pitch angle. These parameters together describe the spatial attitude of the UAV relative to each key monitoring point.

[0042] The relative distance between the current location of the drone and each preset key monitoring point is compared with the optimal observation altitude range, and the relative angle between the current location of the drone and each preset key monitoring point is compared with the optimal observation angle range.

[0043] The calculated relative distance is compared with the upper and lower limits of the preset optimal observation altitude range, and the relative angle is compared with the upper and lower limits of the preset optimal observation angle range. Only when both the relative distance and relative angle of the UAV fall within their respective optimal observation ranges is the UAV considered to have met the preliminary conditions for data collection at the key monitoring point.

[0044] If the relative distance between the current location of the drone and one or more preset key monitoring points exceeds the optimal observation height range, or the relative angle between the current location of the drone and one or more preset key monitoring points exceeds the optimal observation angle range, then the drone is determined not to meet the data collection requirements of the key monitoring points.

[0045] If the distance or angle between the drone and any key monitoring point does not meet the preset optimal observation conditions, then the current location is determined to be unsuitable for high-quality data collection.

[0046] The system identifies key monitoring points where the drone does not meet data collection requirements, and also identifies the deviation between the drone's current location and these key monitoring points.

[0047] After completing the aforementioned determination, record which key monitoring points could not be effectively collected, and record in detail the specific deviation between the UAV's current position and the key monitoring points that do not meet the requirements. For example, how many meters higher / lower the distance, how many degrees left / right the angle, etc.; the recorded deviation information is of great guiding significance for subsequent path planning and adjustments.

[0048] S120: If it is determined that the UAV does not meet the data collection requirements of the key monitoring points, the environmental perception data around the UAV is collected through Beidou navigation, and based on the environmental perception data, a safe flight area for the UAV is constructed within the spatial range centered on the UAV.

[0049] In embodiments of the present invention, environmental perception data refers to the surrounding environment information acquired by the UAV through its onboard sensors (such as radar, visual sensors, etc.), used to identify obstacles, assess flight risks, etc. The safe flight zone of the UAV refers to the spatial range within which the UAV can safely fly under the current environmental perception data, avoiding identified obstacles.

[0050] The environmental perception data collected in this method can be achieved through various sensors carried by the UAV. For example, the UAV can be equipped with LiDAR, which emits laser beams and receives reflected signals to acquire three-dimensional point cloud data of the surrounding environment, thereby identifying the location and size of obstacles. In addition, the UAV can also be equipped with visual sensors, which use image processing algorithms to identify obstacles in the environment, such as trees, buildings, and high-voltage lines. The BeiDou Navigation Satellite System can provide high-precision position and attitude information of the UAV during this process, assisting in the accurate matching of environmental perception data with the actual geographical location. Based on this environmental perception data, a three-dimensional environmental model can be constructed within a spatial range centered on the UAV, and all obstacles can be identified within this model. The construction of the UAV's safe flight zone involves excluding the space occupied by all obstacles within this three-dimensional environmental model, thus obtaining a collision-free area where the UAV can fly freely.

[0051] Specifically, this method collects environmental perception data around the UAV using BeiDou navigation. It utilizes various sensors onboard the UAV, such as lidar, visual sensors, and ultrasonic sensors, combined with high-precision positioning information provided by the BeiDou navigation system, to perform real-time detection and data collection of the environment surrounding the UAV's current location. Environmental perception data includes, but is not limited to, distance information, image information, and point cloud data. Its purpose is to comprehensively understand the potential risks along the UAV's flight path.

[0052] Obtaining obstacle information within the space centered on the drone involves identifying and extracting entities that pose a potential threat to the drone's flight from collected environmental perception data through data processing and analysis algorithms. These entities include buildings, trees, power lines, bridges, other aircraft, and birds. Obstacle information typically includes parameters such as the obstacle's location, size, shape, and relative distance to the drone.

[0053] Based on environmental perception data surrounding the drone and obstacle information within the drone's spatial range, a safe flight zone for the drone is constructed. This involves identifying obstacles and, according to a preset safety margin, delineating one or more areas within the space surrounding the drone where it can fly safely. This safe flight zone is dynamically adjusted; its boundaries change based on real-time obstacle information to ensure the drone avoids collisions during mission execution. For example, obstacle distribution can be represented using 3D grid maps, octree maps, or point cloud maps. Based on this, path planning or obstacle avoidance algorithms are used to calculate the free space where the drone can safely pass, i.e., the safe flight zone.

[0054] By clearly defining and refining the construction process of the UAV safe flight zone, the UAV can more accurately and reliably identify obstacles in its surrounding environment when it determines that the data collection requirements of key monitoring points are not met. Specifically, the BeiDou Navigation Satellite System is first used to assist the UAV's onboard sensors in collecting comprehensive environmental perception data, ensuring the accuracy and real-time nature of data collection. Subsequently, key obstacle information is extracted from this perception data, providing the necessary foundation for the subsequent construction of the safe flight zone. Finally, based on this accurate obstacle information, a reliable safe flight zone for the UAV can be constructed, thus providing a collision-free flight space for the UAV to adjust its flight path to meet data collection requirements.

[0055] S130: Based on key monitoring point information and the safe flight area of ​​UAVs, calculate temporary monitoring points that enable the UAV's real-time location to meet the data collection requirements of key monitoring points, and construct the river and lake patrol flight path between the UAV and the temporary monitoring points.

[0056] In embodiments of the present invention, a temporary monitoring point refers to an alternative point calculated within a safe flight area when the UAV cannot directly meet the data collection requirements of the key monitoring point. This alternative point enables the UAV to meet the data collection requirements. The river and lake patrol flight path refers to the trajectory of the UAV from its current location to the temporary monitoring point, ultimately completing the data collection task.

[0057] Specifically, within the safe flight area of ​​the drone, path planning algorithms can be used to find one or more temporary monitoring points. For example, starting from the drone's current position, sampling points can be randomly set within the safe flight area, and the nearest critical monitoring point can be found among the critical monitoring points. Then, a temporary monitoring point is generated by extending a preset step length from the nearest critical monitoring point towards the sampling point. After generating the temporary monitoring point, it is necessary to check whether the drone meets the data collection requirements when using the temporary monitoring point as a critical monitoring point. If it does, a river and lake patrol flight path is established from the drone's current position to the temporary monitoring point based on the temporary monitoring point. If it does not meet the requirements, new sampling points are set and the aforementioned steps are executed until a temporary monitoring point that meets the data collection requirements of the critical monitoring point is obtained. The flight path can be constructed using algorithms such as A* and RRT (Fast Random Tree), which can generate an optimal or suboptimal path from the drone's current position to the temporary monitoring point, taking into account obstacles and flight constraints.

[0058] In the above implementation, although a scheme for calculating temporary monitoring points and constructing flight paths based on key monitoring point information and the safe flight area of ​​UAVs is proposed, in practical applications, how to efficiently and safely determine these temporary monitoring points and ensure that the constructed flight paths can effectively avoid obstacles and meet data collection requirements is still a technical problem that needs further refinement and optimization.

[0059] This application further proposes a method for constructing a river and lake patrol flight path between the drone and the temporary monitoring points, based on the key monitoring point information and the safe flight area of ​​the drone, to calculate temporary monitoring points that ensure the real-time location of the drone meets the data collection requirements of the key monitoring points. The method includes: Starting from the drone's current location, sampling points are randomly set within the drone's safe flight area, and the nearest key monitoring point is found among the key monitoring points.

[0060] Starting from the drone's current location, sampling points are randomly selected within the drone's safe flight area. This aims to expand the search range for potential temporary monitoring points through random exploration, while ensuring the drone's flight safety. The safe flight area refers to the designated space for safe flight of the drone, taking into account obstacle information, no-fly zones, and the drone's own performance limitations. The randomly selected sampling points can be understood as several candidate locations randomly chosen within the safe flight area to guide the subsequent generation of temporary monitoring points.

[0061] Furthermore, among the key monitoring points, the nearest key monitoring point to the sampling point is found. This aims to ensure a certain correlation between the generated temporary monitoring point and the original key monitoring points, thereby guaranteeing the effectiveness of the data acquisition task. The nearest key monitoring point can be determined by calculating the Euclidean distance or other suitable distance metric between the sampling point and all key monitoring points.

[0062] Temporary monitoring points are generated by extending a preset step length from the nearest key monitoring point toward the sampling point.

[0063] Starting from the nearest critical monitoring point, extend along the direction pointing to the sampling point by a preset fixed or variable length to obtain a new location point, i.e., a temporary monitoring point. The preset step length can be set according to factors such as the actual application scenario, the drone's flight speed, and the sensor's observation range. The purpose is to ensure the correlation between the temporary monitoring point and the critical monitoring point while allowing it to more flexibly adapt to the limitations of the drone's current location and safe flight area.

[0064] When the temporary monitoring point is used as a key monitoring point, check whether the UAV meets the data collection requirements of the key monitoring point.

[0065] The evaluation process assesses whether the drone at the temporary monitoring point meets the preset requirements for key monitoring points, including geographical coordinates, optimal observation altitude range, and optimal observation angle range. If these requirements are met, the temporary monitoring point is considered valid and can be used as a new data collection target for the drone. If not, the temporary monitoring point is unsuitable for data collection and needs to be relocated.

[0066] If the conditions are met, a river and lake patrol flight path is established from the current location of the UAV to the temporary monitoring point based on the temporary monitoring point; if the conditions are not met, new sampling points are set and the aforementioned steps are executed until a temporary monitoring point that meets the data collection requirements of the key monitoring point is obtained.

[0067] If the data collection requirements are met, a river and lake patrol flight path is established from the UAV's current location to the temporary monitoring point, based on the temporary monitoring point. This flight path can be established using path planning algorithms such as A* algorithm or RRT (Fast Random Tree) algorithm. Considering factors such as obstacle avoidance and energy consumption optimization within the UAV's safe flight area, a collision-free, optimal, or suboptimal flight trajectory from the UAV's current location to the temporary monitoring point is generated. If the requirements are not met, new sampling points are set, and the aforementioned steps are executed until a temporary monitoring point that meets the data collection requirements of the key monitoring point is obtained. This iterative search mechanism ensures that even if the initial attempt fails to find a suitable temporary monitoring point, the system can eventually find a solution that meets the conditions through multiple attempts.

[0068] The following is an illustration using a specific example. Assume a drone is patrolling a section of a river, currently located at point A. The system's judgment module detects that point A does not meet the data acquisition requirements for a nearby key monitoring point K1 (e.g., too far from K1 or poor observation angle). At this point, the drone's safe flight area construction module has already constructed a safe flight area centered on the drone based on environmental perception data. The flight path construction module will then initiate the temporary monitoring point calculation process. A sampling point S1 is randomly selected within the drone's safe flight area. Among all key monitoring points (assuming K1, K2, and K3), K1 is calculated as the key monitoring point closest to S1. Next, a preset step size (e.g., 0.5 meters) is extended from K1 towards S1, generating a temporary monitoring point T1. The system then checks whether the drone meets the data acquisition requirements when T1 is used as the key monitoring point. If T1 still does not meet the requirements (e.g., although it is closer to K1, the observation angle is still not ideal), continue to randomly set new sampling points S2 within the safe flight area and repeat the above steps to find the key monitoring point closest to S2, and extend it to generate new temporary monitoring points T2, until a temporary monitoring point T_final that meets the data collection requirements is found. Once T_final is found, based on T_final, establish a river and lake patrol flight path from the UAV's current position A to T_final. This path will take into account obstacle avoidance within the safe flight area to ensure that the UAV can safely reach T_final and complete data collection.

[0069] In an embodiment of this invention, a scheme for constructing a flight path for river and lake patrols between a drone and a temporary monitoring point is proposed. However, in actual river and lake patrols, the drone may face dynamic changes in environmental factors such as light intensity, distance from the monitoring point, and flight speed while flying along the preset flight path. These dynamic changes may lead to a decrease in the quality of data collected by the airborne sensors, such as blurred images, inaccurate exposure, or distorted spectral data, thereby affecting the effectiveness of the patrol task and the accuracy of data analysis. If the above problems are not solved, even if the flight path is properly planned, the final acquired data may not meet the requirements of high-quality monitoring, requiring additional manual intervention or repeated flights, thus reducing patrol efficiency. To address this, this application further proposes a step of optimizing the flight path by adaptively adjusting the airborne sensor parameters of the drone after the step of constructing the flight path between the drone and the temporary monitoring point, aiming to ensure the acquisition of high-quality monitoring data even in dynamic environments.

[0070] Specifically, after constructing the flight path between the drone and the temporary monitoring point, the process also includes optimizing the flight path by adaptively adjusting the drone's onboard sensor parameters, including: Focal length adjustment: If the distance between the UAV and a temporary monitoring point in the flight path is greater than a first preset distance, the focal length parameter of the UAV's onboard image acquisition sensor will be adaptively adjusted. The first preset distance can be set based on the optical characteristics of the image acquisition sensor and the required image resolution. When the distance between the UAV and the monitoring point is far, increasing the focal length allows for magnification of the target area, thereby acquiring clearer and more detailed images of the monitoring point without changing the UAV's flight altitude. Improving the clarity and detail of images of key monitoring points is beneficial for flight path planning.

[0071] Exposure Adjustment: If the light intensity changes during the drone's flight path, the exposure time and sensitivity of the drone's onboard image acquisition sensor will be adaptively adjusted. Changes in light intensity may be caused by factors such as cloud cover, the flight area entering shadow, or direct sunlight. By monitoring the ambient light intensity in real time and adjusting the exposure time and sensitivity accordingly, overexposure or underexposure of the image can be avoided, ensuring that the image has appropriate brightness and contrast.

[0072] Frequency Adjustment: If the drone's flight speed increases or the sampling frequency rises, the scanning frequency of the multispectral sensor will be adaptively adjusted. When the drone's flight speed increases, the scanning frequency of the multispectral sensor needs to be increased to ensure that the area covered per unit time is fully scanned. Similarly, if there are higher requirements for the sampling frequency of data acquisition, the scanning frequency also needs to be adjusted accordingly to ensure that sufficient density of spectral data is obtained.

[0073] Exposure time adjustment for different spectral bands: The exposure time for each spectral band is adaptively adjusted to address the varying sensitivities of different spectral bands to light intensity. For example, some spectral bands may be more sensitive to specific wavelengths of light, or more prone to saturation or underexposure under certain lighting conditions. By independently adjusting the exposure time for each spectral band, the data acquisition effect of different bands can be optimized, improving the overall quality and accuracy of multispectral data.

[0074] Image quality adjustment: The image quality acquired by the image acquisition sensor is evaluated. If the image sharpness or spectral data signal-to-noise ratio does not meet preset requirements, the shooting parameters of the image acquisition sensor are adaptively fine-tuned. Image quality evaluation can be performed by calculating indicators such as image sharpness, contrast, and signal-to-noise ratio. When the evaluation result is lower than a preset threshold, the system can automatically fine-tune shooting parameters such as focal length, aperture, and shutter speed to achieve optimal image sharpness and spectral data signal-to-noise ratio, ensuring the usability of the acquired data.

[0075] This embodiment introduces an adaptive adjustment mechanism for onboard sensor parameters after the UAV flight path is constructed, effectively solving the data quality assurance problem of traditional path planning methods in dynamic environments. Specifically, during UAV patrols of rivers and lakes, factors such as the distance between the UAV and temporary monitoring points, ambient light intensity, flight speed, and data sampling frequency requirements constantly change. These changes directly affect the working status and data output quality of the image acquisition sensor and multispectral sensor. By monitoring these environmental and task parameters in real time and dynamically adjusting the sensor's focal length, exposure time, sensitivity, scanning frequency, and exposure time for different spectral bands according to preset rules or algorithms, the sensors can always operate in optimal condition. For example, when the UAV is far from the monitoring point, adaptive adjustment of the focal length ensures clear imaging of distant targets; when the lighting changes, adjustment of exposure parameters avoids overexposure or underexposure; when the flight speed increases or sampling requirements rise, adjustment of the scanning frequency ensures the integrity of data coverage. In addition, real-time evaluation and fine-tuning of image sharpness and spectral data signal-to-noise ratio provide a closed-loop feedback mechanism, further ensuring that the quality of the final acquired data meets the requirements. It is precisely because of this real-time, multi-dimensional adaptive adjustment of sensor parameters that drones can continuously acquire high-quality, high-reliability monitoring data in complex river and lake patrol environments.

[0076] In practical applications, simply adjusting parameters such as focal length, exposure, and frequency may not be sufficient to handle complex and varied lighting environments, leading to localized overexposure or underexposure in images of key monitoring points, thus affecting the quality and accuracy of data acquisition. This embodiment adaptively adjusts the image acquisition sensor's shooting parameters to obtain optimal images of key monitoring points, specifically including: Illumination difference detection is performed on key monitoring point images acquired by image acquisition sensors, and abnormal illumination areas in the key monitoring point images are identified based on the brightness difference detection results; among them, abnormal illumination areas include overexposed areas and underexposed areas.

[0077] In this embodiment, illumination difference detection analyzes the pixel brightness distribution of an image, using methods such as histogram analysis and local brightness statistics to identify areas in the image whose brightness values ​​exceed a preset threshold range, thus classifying them as areas of abnormal illumination. Overexposed areas refer to regions in the image with excessively high brightness, resulting in loss of detail; underexposed areas refer to regions in the image with excessively low brightness, resulting in blurred or invisible details. The aim is to accurately identify data quality problems in the image caused by uneven illumination.

[0078] By fine-tuning the drone's flight attitude parameters, the shooting angle and / or shooting distance of the image acquisition sensor can be adaptively adjusted.

[0079] Fine-tuning flight attitude parameters refers to making small-scale, precise adjustments to the pitch, roll, and yaw of a UAV. These adjustments can change the shooting angle of the image acquisition sensor relative to key monitoring points, thereby altering the angle at which light strikes the sensor, or changing the shooting distance to optimize image lighting and avoid or mitigate localized lighting anomalies.

[0080] After each fine-tuning of the flight attitude parameters, the UAV continuously acquires multiple images of key monitoring points based on the image acquisition sensor. All images of key monitoring points form a key monitoring point image sequence. The key monitoring point image sequence contains multiple images of the same area under different lighting angles.

[0081] The key monitoring point image sequence is aligned and analyzed to obtain the optimal key monitoring point image.

[0082] Continuously acquiring multiple images ensures that scene information under different lighting conditions can be captured after attitude fine-tuning, forming an image set rich in lighting information. These image sequences provide the data foundation for subsequent lighting optimization. Alignment refers to eliminating image displacement caused by UAV attitude fine-tuning through image registration technology, ensuring that the same scene area in the image sequence can correspond accurately. Analyzing the key monitoring point image sequence aims to extract the best visual information from multiple images, such as by fusing images with different exposures to generate an image with good detail and color representation in both bright and dark areas, thereby obtaining the optimal key monitoring point image.

[0083] Specifically, the steps of aligning and analyzing the key monitoring point image sequence to obtain the optimal key monitoring point image may further include the following operations: A feature-point-based image registration algorithm is used to align the image sequence of the key monitoring points to compensate for the image displacement caused by fine-tuning the UAV's attitude.

[0084] A feature-point-based image registration algorithm is used to align key monitoring point image sequences. This involves identifying unique feature points in the images and utilizing the correspondence between these feature points in different images to calculate geometric transformation parameters, thereby precisely aligning all images in the sequence to the same coordinate system. This approach aims to effectively compensate for image displacements that may result from fine-tuning of the UAV's flight attitude, ensuring accurate overlap of different images of the same scene within the key monitoring point image sequence, thus laying the foundation for subsequent image fusion. Commonly used feature-point algorithms include Scale Invariant Feature Transform (SIFT), Speed-Up Robust Feature Transform (SURF), and Oriented Fast and Rotationally Short Feature Transform (ORB).

[0085] The aligned key monitoring point image sequence is analyzed using a multi-exposure image fusion algorithm. One key monitoring point image is selected from the aligned key monitoring point image sequence as the initial optimal key monitoring point image.

[0086] Multiple images acquired under different exposure conditions are intelligently combined to generate an image with a wider dynamic range and richer details. Specifically, firstly, a key monitoring point image is selected from the aligned key monitoring point image sequence as the initial optimal key monitoring point image. This image is typically one with relatively centered exposure or good overall visual effect. Subsequently, targeted processing is performed on the abnormal lighting areas identified in the initial optimal key monitoring point image, namely overexposed and underexposed areas.

[0087] For overexposed areas in the initial optimal key monitoring point image, select identical areas with weak reflection and clear details from other key monitoring point images in the aligned key monitoring point image sequence (excluding the initial optimal key monitoring point image) for replacement. For underexposed areas in the initial optimal key monitoring point image, select identical areas with high brightness and obvious texture from other key monitoring point images in the key monitoring point image sequence (excluding the initial optimal key monitoring point image) for replacement. The key monitoring point image obtained after replacement is used as the optimal key monitoring point image.

[0088] Overexposed areas typically lose detail due to excessive light, while underexposed images may retain more detail in those areas. Conversely, for underexposed areas in the initial optimal key monitoring point image, similar areas with high brightness and clear texture are selected from other key monitoring point images in the image sequence for replacement. This is because underexposed areas are usually too dark due to insufficient light, while other fully exposed images may provide richer brightness information and texture details in those areas. Through this regional intelligent replacement, the final key monitoring point image after replacement is selected as the optimal key monitoring point image.

[0089] In actual river and lake patrol missions, in addition to image or spatial data acquisition, water quality parameter monitoring is equally crucial. Traditional drone water quality monitoring may only perform simple data collection and transmission, lacking intelligent judgment and processing mechanisms for real-time changes in water quality data. It is easily affected by environmental noise or instantaneous fluctuations, leading to low data transmission efficiency or misjudgment of water quality conditions.

[0090] In response, this application further proposes a scheme for adaptively adjusting the parameters of the water quality sensor, specifically including: When the drone receives a water quality data collection command, it controls the water quality sensor to simultaneously collect water quality parameters in the water body at key monitoring points. The types of water quality parameters include pH value, conductivity, turbidity, and dissolved oxygen content. Real-time monitoring of each type of water quality parameter, statistical analysis of the instantaneous change rate of each type of water quality parameter, and determination of whether there are abnormal fluctuations in the instantaneous change rate of each type of water quality parameter. If the instantaneous change rate of two or fewer types of water quality parameters fluctuates abnormally at the same time, it is determined that the water quality at the key monitoring point has not changed substantially. The water quality parameters are then calibrated and filtered, and the calibrated and filtered water quality parameters are transmitted to the host system through the communication system of the drone. If the instantaneous change rate of two or more types of water quality parameters fluctuates abnormally at the same time, and it is determined that the water quality at the key monitoring point has undergone substantial changes, then the raw data of the water quality parameters will be transmitted to the host system through the communication system of the drone.

[0091] Specifically, adaptive adjustment of water quality sensor parameters refers to the UAV intelligently adjusting the configuration and operating mode of its onboard water quality sensors based on actual monitoring needs and environmental conditions to optimize the quality and efficiency of water quality data acquisition. Water quality data acquisition commands can be issued by a host system or triggered by the UAV based on preset tasks or its own status. Upon receiving the command, the UAV will activate its onboard water quality sensors, extending the sensor probes into the water body above key monitoring points or at a specific height to simultaneously collect various water quality parameters. These water quality parameters are key indicators for assessing the health of water bodies. For example, pH reflects the acidity or alkalinity of the water, conductivity is related to the concentration of dissolved ions in the water, turbidity indicates the degree of turbidity, and dissolved oxygen content is an important condition for the survival of aquatic organisms.

[0092] This system involves real-time monitoring of each type of water quality parameter and statistical analysis of its instantaneous change rate, aiming to capture the dynamic trends of these parameters. The instantaneous change rate can be calculated using continuously collected water quality data points, for example, by comparing the difference between the current sample value and the previous sample value and dividing by the sampling time interval. The presence of abnormal fluctuations can be determined based on preset thresholds or statistical models. For instance, a normal fluctuation range can be set, and any instantaneous change rate exceeding this range is considered an abnormal fluctuation.

[0093] In practical applications, the logic for handling abnormal fluctuations is the core of this solution. If only two or fewer types of water quality parameters exhibit abnormal fluctuations in their instantaneous rate of change simultaneously, this is usually considered a local or minor environmental disturbance, such as a sensor probe briefly contacting floating objects on the water surface, or the instantaneous effect of local water flow, rather than a substantial change in the overall water quality. In this case, to ensure the accuracy and reliability of the data, the collected water quality parameters will undergo data calibration and filtering. Data calibration may include eliminating sensor drift and temperature compensation, while filtering may employ algorithms such as moving average and median filtering to remove noise. The processed water quality parameters are considered reliable and are transmitted to the host system via the UAV's communication system for further analysis and storage.

[0094] However, if the instantaneous change rates of two or more types of water quality parameters exhibit abnormal fluctuations simultaneously, this strongly indicates that the water quality at key monitoring points may have undergone substantial changes, such as being affected by pollutant discharge or large-scale ecological events. In such cases, to preserve the original, unprocessed on-site information, the raw water quality parameter data will be directly transmitted to the higher-level system via the drone's communication system. The purpose of this is to allow the higher-level system to obtain the most original and comprehensive data, enabling experts to conduct in-depth analysis and judgment, and thus take timely countermeasures.

[0095] In some preferred embodiments, a specific example is given below. Suppose a drone is conducting a routine patrol of a river area. When the drone flies over a pre-set key monitoring point, it receives a water quality data collection command from the host system. The drone then controls its onboard water quality sensor probe to extend into the water body and begin simultaneously collecting the pH, conductivity, turbidity, and dissolved oxygen content at that key monitoring point.

[0096] During the data collection process, the drone monitors the instantaneous change rates of these four water quality parameters in real time. For example, at a certain moment, the system detects a sudden increase in the instantaneous change rate of the turbidity parameter, while the instantaneous change rates of the other three parameters (pH, conductivity, and dissolved oxygen content) remain within the normal range. According to the judgment logic of this scheme, since only one water quality parameter (turbidity) shows abnormal fluctuations in its instantaneous change rate (i.e., fewer than two), the system determines that the current water quality has not undergone substantial changes. At this time, the drone will perform data calibration (e.g., correction based on sensor calibration curves) and filtering on the collected water quality data (e.g., using moving average filtering to remove instantaneous noise), and then transmit the processed, more reliable water quality parameters to the host system through its communication system.

[0097] For example, at another moment, when the drone is collecting water quality data at another key monitoring point, it simultaneously detects significant abnormal fluctuations in the instantaneous change rates of pH, conductivity, and dissolved oxygen content, while the instantaneous change rate of turbidity also slightly increases. According to the logic of this solution, since the instantaneous change rates of three water quality parameters (pH, conductivity, and dissolved oxygen content) are simultaneously exhibiting abnormal fluctuations (i.e., more than two), the system determines that the water quality at this key monitoring point may have undergone substantial changes. In this case, the drone will directly transmit all the raw data of the collected water quality parameters, including uncalibrated and unfiltered raw values, to the host system via its communication system. Upon receiving this raw data, the host system can immediately initiate higher-level analysis procedures and even notify relevant departments to conduct on-site verification in order to promptly identify and address potential water pollution problems.

[0098] In some embodiments described above, this application proposes a scheme to optimize flight paths by adaptively adjusting the parameters of the UAV's onboard sensors and further to obtain optimal images of key monitoring points by adjusting the image acquisition sensor parameters. However, in actual river and lake patrol missions, UAVs may encounter environments with low visibility, such as fog, rain, snow, or nighttime operations. In such low-visibility environments, traditional image acquisition and path planning methods may face challenges, leading to decreased image quality, inaccurate positioning, and consequently affecting the data acquisition effect and patrol efficiency of key monitoring points.

[0099] In response, this application further proposes a scheme that includes the following steps before the step of adaptively adjusting the image acquisition sensor's shooting parameters: The drone's onboard environmental perception sensors collect environmental visibility data of the drone's location and compare it with a preset visibility threshold.

[0100] The onboard environmental perception sensors of a drone can collect environmental visibility data of the drone's location in real time or periodically. These sensors can be visible light sensors, infrared sensors, or lidar, used to measure visibility-related parameters such as air transparency or obstacle density. The collected visibility data is compared with a preset visibility threshold to determine whether the current environment is considered low visibility. This visibility threshold can be set according to the actual application scenario and the performance of the drone's sensors; for example, visibility below 50 meters can be considered low visibility.

[0101] When the environmental visibility data collected by the environmental perception sensor is lower than the visibility threshold, the drone performs a rough location determination of the current location and combines it with the historical map information stored in the drone's onboard system to determine the prior scene structure that best matches the drone's current location.

[0102] When the visibility data collected by the environmental perception sensors is below the visibility threshold, it indicates that the UAV is in a low-visibility environment, requiring special positioning and environmental modeling strategies. The UAV performs a rough location assessment, which can be achieved by combining location information provided by the BeiDou Navigation Satellite System with inertial measurement unit (IMU) data. Simultaneously, it utilizes historical map information stored within the UAV's onboard system, such as high-precision geographic information system (GIS) data, 3D point cloud maps, or pre-built scene models, to determine the prior scene structure that best matches the UAV's current location. This prior scene structure can be a predefined building model, terrain model, or river structure model, providing an initial reference for subsequent precise positioning and environmental perception.

[0103] The geometric parameters of the prior scene structure are adjusted and updated by using sparse point cloud data collected by UAV-borne LiDAR and blurred visual images collected by image acquisition sensors under low visibility conditions.

[0104] To accurately adjust and update the prior scene structure under low visibility conditions, this application utilizes multi-source heterogeneous sensor data. Sparse point cloud data acquired by an UAV-borne LiDAR can obtain depth information and geometric structure of the environment, providing relatively reliable distance measurements even in low visibility. Simultaneously, blurred visual images acquired by image acquisition sensors under low visibility, while having limited sharpness, still contain texture and contour information. These sparse point cloud data and blurred visual images are fused and processed to adjust and update the geometric parameters of the prior scene structure. For example, a point cloud registration algorithm can be used to match the real-time point cloud with the prior scene structure, and the blurred visual image can be used to assist in correcting the matching result, thereby updating the geometric parameters such as size, position, or orientation in the prior scene structure.

[0105] Based on the adjustment and update of the geometric parameters of the prior scene structure, an environmental topology graph of the prior scene structure is constructed; where nodes in the environmental topology graph represent the central region of key monitoring points, and edges in the environmental topology graph represent feasible paths between nodes.

[0106] Based on the adjustment and updating of the geometric parameters of the prior scene structure, an environmental topology map of the prior scene structure can be constructed. This environmental topology map is an abstract representation of the environment, where nodes represent the central areas of key monitoring points, such as specific sampling points in a river, under a bridge, or near a sluice gate. These nodes typically have explicit geographic coordinates and semantic information. The edges in the environmental topology map represent feasible paths between nodes; these paths are connections that have undergone safety assessments and are suitable for UAV flight. By constructing the environmental topology map, complex environmental information can be simplified into a structure that is easy to plan and navigate, providing an efficient decision-making basis for UAV path planning in low visibility conditions.

[0107] This application's solution effectively addresses the limitations of traditional methods in adverse weather or lighting conditions by introducing a perception and modeling mechanism for low-visibility environments before adaptive adjustment of image acquisition sensor parameters. During UAV river and lake patrols, especially in low-visibility environments, this application significantly enhances the UAV's environmental perception capabilities and path planning robustness. This solution avoids relying directly on potentially distorted visual information for decision-making in low-visibility conditions. Instead, it achieves accurate environmental modeling and topology construction through multi-sensor fusion and the utilization of prior knowledge. Therefore, even in adverse conditions such as fog, rain, snow, or nighttime, the UAV can accurately identify the location of key monitoring points and plan safe and efficient flight paths, ensuring the quality of data acquisition and the smooth progress of patrol tasks. This not only improves the environmental adaptability of the UAV system but also provides more reliable input for subsequent adaptive adjustment of image acquisition sensor parameters, ultimately improving the overall efficiency and data accuracy of river and lake patrols.

[0108] In some preferred embodiments, it is assumed that a drone is performing a nighttime river and lake patrol mission under low visibility conditions. First, the drone's onboard environmental perception sensors (e.g., a laser rangefinder combined with a photoelectric sensor) continuously collect environmental visibility data. When the detected visibility data falls below a preset nighttime visibility threshold (e.g., 50 meters), a perception and planning process for low-visibility environments is initiated. The drone uses its BeiDou navigation module and inertial measurement unit for coarse positioning and retrieves a 3D model of the bridge structure in the area from its onboard stored historical map information as a priori scene structure. The drone's onboard LiDAR begins collecting sparse point cloud data, while image acquisition sensors acquire blurred visual images under low-light conditions. A point cloud registration algorithm matches the real-time point cloud with the bridge structure model, and the blurred visual images are used to assist in identifying the bridge's outline structure, thereby adjusting and updating the geometric parameters of the bridge model to better match the current actual environment. Based on the updated bridge model, an environmental topology graph is constructed, where specific sampling points of the bridge are defined as nodes, and flight paths between these nodes, after safety assessment, are defined as edges. For example, if a drone needs to fly from node A to node B, the system will plan a feasible path based on the topology map. After constructing the environmental topology map, the drone can use this map, combined with key monitoring point information, to calculate temporary monitoring points that meet the data acquisition requirements, and construct a river and lake patrol flight path from its current location to the temporary monitoring points. Based on this, the image acquisition sensor's shooting parameters are adaptively adjusted to obtain optimal images of the key monitoring points.

[0109] like Figure 2 As shown, the present invention also provides a UAV river and lake patrol path planning system 200 based on Beidou navigation, comprising: The judgment module 210 is used to acquire the real-time location information of the UAV and, based on the preset key monitoring point information and the real-time location information of the UAV, determine whether the UAV meets the data collection requirements of the key monitoring points. The UAV safe flight area construction module 220 is used to collect environmental perception data around the UAV through Beidou navigation when it is determined that the UAV does not meet the data collection requirements of key monitoring points, and to construct a safe flight area of ​​the UAV within a spatial range centered on the UAV based on the environmental perception data. The flight path construction module 230 is used to calculate temporary monitoring points that meet the data collection requirements of key monitoring points based on key monitoring point information and the safe flight area of ​​UAVs, and to construct the river and lake patrol flight path between the UAV and the temporary monitoring points.

[0110] Although embodiments of the present invention have been shown and described above, it is understood that the embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the embodiments within the scope of the present invention.

Claims

1. A method for unmanned aerial vehicle (UAV) river and lake patrol path planning based on BeiDou navigation, characterized in that, include: The system acquires the real-time location information of the drone and, based on the preset key monitoring point information and the real-time location information of the drone, determines whether the drone meets the data collection requirements of the key monitoring points. If it is determined that the UAV does not meet the data collection requirements of the key monitoring points, environmental perception data around the UAV is collected through Beidou navigation, and based on the environmental perception data, a safe flight area for the UAV is constructed within a spatial range centered on the UAV. Based on the key monitoring point information and the safe flight area of ​​the UAV, a temporary monitoring point is calculated to ensure that the real-time position of the UAV meets the data collection requirements of the key monitoring point, and a river and lake patrol flight path is constructed between the UAV and the temporary monitoring point.

2. The method for unmanned aerial vehicle (UAV) river and lake patrol path planning based on BeiDou navigation as described in claim 1, characterized in that, Acquire real-time location information of the drone, and based on preset key monitoring point information and the real-time location information of the drone, determine whether the drone meets the data collection requirements of the key monitoring points, including: Obtain the real-time location information of the drone, either in real time or at regular intervals. Obtain the preset key monitoring point information, wherein the preset key monitoring point information includes the geographic coordinate information, optimal observation height range, and optimal observation angle range of all key monitoring points in the patrol task; Based on the geographic coordinates of all key monitoring points and the real-time location of the UAV, the relative distance and relative angle between the current location of the UAV and each preset key monitoring point are calculated. The relative distance between the current location of the UAV and each preset key monitoring point is compared with the optimal observation height range, and the relative angle between the current location of the UAV and each preset key monitoring point is compared with the optimal observation angle range. If the relative distance between the current location of the UAV and one or more preset key monitoring points exceeds the optimal observation height range, or the relative angle between the current location of the UAV and one or more preset key monitoring points exceeds the optimal observation angle range, then the UAV is determined to not meet the data collection requirements of the key monitoring points. The system identifies the key monitoring points where the drone does not meet the data collection requirements, and also identifies the deviation between the drone's current location and the key monitoring points that do not meet the data collection requirements.

3. The unmanned aerial vehicle river and lake patrol path planning method based on Beidou navigation according to claim 1, characterized in that, The system collects environmental perception data around the UAV using BeiDou navigation, and based on this data, constructs a safe flight zone for the UAV within a spatial range centered on the UAV, including: The BeiDou navigation system collects environmental perception data around the drone to obtain obstacle information within the spatial range centered on the drone. Based on the environmental perception data around the drone and the obstacle information within the spatial range centered on the drone, a safe flight zone for the drone is constructed within the spatial range centered on the drone.

4. The unmanned aerial vehicle river and lake patrol path planning method based on Beidou navigation according to claim 1, characterized in that, Based on the key monitoring point information and the safe flight area of ​​the UAV, temporary monitoring points are calculated to ensure that the real-time position of the UAV meets the data collection requirements of the key monitoring points. A river and lake patrol flight path is then constructed between the UAV and the temporary monitoring points, including: Starting from the current position of the UAV, sampling points are randomly set within the safe flight area of ​​the UAV, and the nearest key monitoring point is found among the key monitoring points that is closest to the sampling point. A temporary monitoring point is generated by extending a preset step length from the nearest key monitoring point toward the sampling point; When the temporary monitoring point is used as a key monitoring point, check whether the UAV meets the data collection requirements of the key monitoring point; If the conditions are met, a river and lake patrol flight path is established from the current location of the UAV to the temporary monitoring point based on the temporary monitoring point; if the conditions are not met, new sampling points are set and temporary monitoring points are generated until a temporary monitoring point that meets the data collection requirements of the key monitoring point is obtained.

5. The unmanned aerial vehicle river and lake patrol path planning method based on Beidou navigation according to claim 4, characterized in that, Following the step of constructing the flight path between the UAV and the temporary monitoring point, the method further includes a step of optimizing the flight path by adaptively adjusting the onboard sensor parameters of the UAV, including: Focal length adjustment: If the distance between the UAV and the temporary monitoring point in the flight path is greater than the first preset distance, the focal length parameter of the UAV's onboard image acquisition sensor will be adaptively adjusted. Exposure adjustment: If the light intensity changes during the flight of the UAV along the flight path, the exposure time and sensitivity of the UAV's onboard image acquisition sensor are adaptively adjusted. Frequency adjustment: If the UAV's flight speed increases or the sampling frequency increases, the scanning frequency of the multispectral sensor will be adaptively adjusted. Exposure time adjustment for different spectral bands: Adaptively adjust the exposure time for each spectral band based on the differences in sensitivity to light intensity among different spectral bands; Image quality adjustment: The image quality acquired by the image acquisition sensor is evaluated. If the image sharpness or spectral data signal-to-noise ratio does not meet the preset requirements, the shooting parameters of the image acquisition sensor are adaptively fine-tuned.

6. The unmanned aerial vehicle river and lake patrol path planning method based on Beidou navigation according to claim 5, characterized in that, The flight path is optimized by adaptively adjusting the onboard sensor parameters of the UAV, and the image acquisition sensor's shooting parameters are also adaptively adjusted to obtain optimal images of key monitoring points, including: Illumination difference detection is performed on the key monitoring point images acquired by the image acquisition sensor, and abnormal illumination regions in the key monitoring point images are identified based on the brightness difference detection results; wherein, the abnormal illumination regions include overexposed regions and underexposed regions; By fine-tuning the flight attitude parameters of the UAV, the shooting angle and / or shooting distance of the image acquisition sensor are adaptively adjusted; After each fine-tuning of the flight attitude parameters of the UAV, multiple images of key monitoring points are continuously acquired based on the image acquisition sensor. All images of key monitoring points are combined into a key monitoring point image sequence. The key monitoring point image sequence contains multiple images of the same area under different lighting angles. The key monitoring point image sequence is aligned and analyzed to obtain the optimal key monitoring point image.

7. The unmanned aerial vehicle river and lake patrol path planning method based on Beidou navigation according to claim 6, characterized in that, Aligning and analyzing the key monitoring point image sequence to obtain the optimal key monitoring point image includes: A feature-point-based image registration algorithm is used to align the image sequence of the key monitoring points to compensate for the image displacement caused by fine-tuning the UAV's attitude. The aligned key monitoring point image sequence is analyzed using a multi-exposure image fusion algorithm. One key monitoring point image is selected from the aligned key monitoring point image sequence as the initial optimal key monitoring point image. For overexposed areas in the initial optimal key monitoring point image, select the same areas with weak reflection and clear details in the key monitoring point images other than the initial optimal key monitoring point image in the aligned key monitoring point image sequence to replace them. For underexposed areas in the initial optimal key monitoring point image, select the same areas with high brightness and obvious texture in key monitoring point images other than the initial optimal key monitoring point image in the key monitoring point image sequence to replace them; The key monitoring point image obtained after substitution is taken as the optimal key monitoring point image.

8. The method for unmanned aerial vehicle (UAV) river and lake patrol path planning based on BeiDou navigation as described in claim 4, characterized in that, The step of optimizing the flight path by adaptively adjusting the onboard sensor parameters of the UAV also includes adaptively adjusting the water quality sensor parameters, including: When the UAV receives a water quality data acquisition command, the UAV controls the water quality sensor to synchronously acquire water quality parameters in the water body at key monitoring points; wherein, the types of water quality parameters include pH value, conductivity, turbidity and dissolved oxygen content; Real-time monitoring of the water quality parameters for each type, statistical analysis of the instantaneous change rate of each type of water quality parameter, and determination of whether there are abnormal fluctuations in the instantaneous change rate of each type of water quality parameter. If the instantaneous change rate of two or fewer types of water quality parameters fluctuates abnormally at the same time, it is determined that the water quality at the key monitoring point has not changed substantially. The water quality parameters are then calibrated and filtered, and the calibrated and filtered water quality parameters are transmitted to the host system through the communication system of the UAV. If the instantaneous change rate of two or more types of water quality parameters fluctuates abnormally at the same time, and it is determined that the water quality at the key monitoring point has undergone substantial changes, then the raw data of the water quality parameters will be transmitted to the host system through the communication system of the drone.

9. A method for planning unmanned aerial vehicle (UAV) river and lake patrol routes based on BeiDou navigation, as described in claim 6, is characterized in that... Before the step of adaptively adjusting the image acquisition sensor's capture parameters, the following steps are also included: The drone's onboard environmental perception sensor collects environmental visibility data of the drone's location and compares it with a preset visibility threshold. When the environmental visibility data collected by the environmental perception sensor is lower than the visibility threshold, the drone performs a rough location determination of the current location and, in conjunction with the historical map information stored in the drone's onboard system, determines the prior scene structure that best matches the drone's current location. The geometric parameters of the prior scene structure are adjusted and updated using sparse point cloud data collected by UAV-borne lidar and blurred visual images collected by image acquisition sensors under low visibility conditions. Based on the adjustment and update of the geometric parameters of the prior scene structure, an environmental topology graph of the prior scene structure is constructed; wherein, the nodes in the environmental topology graph represent the central regions of key monitoring points, and the edges in the environmental topology graph represent feasible paths between nodes.

10. A Beidou navigation-based unmanned aerial vehicle river and lake patrol path planning system, characterized in that, include: The judgment module is used to obtain the real-time location information of the UAV and, based on the preset key monitoring point information and the real-time location information of the UAV, determine whether the UAV meets the data collection requirements of the key monitoring points. The UAV safe flight area construction module is used to collect environmental perception data around the UAV through Beidou navigation when it is determined that the UAV does not meet the data collection requirements of key monitoring points, and to construct a safe flight area for the UAV within a spatial range centered on the UAV based on the environmental perception data. The flight path construction module is used to calculate temporary monitoring points that allow the real-time location of the UAV to meet the data collection requirements of the key monitoring points based on the information of the key monitoring points and the safe flight area of ​​the UAV, and to construct the river and lake patrol flight path between the UAV and the temporary monitoring points.

Images