Water conservancy unmanned aerial vehicle river patrol multi-task intelligent route planning method

Abstract

The present application relates to the technical field of water conservancy unmanned aerial vehicle inspection, and particularly relates to a water conservancy unmanned aerial vehicle river patrol multi-task intelligent flight path planning method, comprising: dividing tasks into three categories of daily, special and emergency and determining priorities; constructing a device, environment and task multi-constraint model, generating differentiated initial flight paths for different tasks and optimizing, and dynamically adjusting the flight path in real time; realizing multi-machine cooperation and cross-airfield task relay; and self-learning and optimizing the flight path based on historical data. The present application solves the problems of poor scene adaptation, lack of multi-task capability and static flight path unable to respond to changes in the prior art, breaks through the endurance limit of a single unmanned aerial vehicle, and greatly improves the efficiency, safety and pertinence of long-distance river channel inspection.

Description

Technical Field

[0001] This invention relates to the technical field of unmanned aerial vehicle (UAV) inspection of water conservancy projects, and in particular to a multi-task intelligent route planning method for UAV river patrol. Background Technology

[0002] With their advantages of maneuverability, wide coverage, low operating costs, and lack of terrain limitations, drones have become an important technical means for river inspection. Flight path planning, as the core link of drone inspection, directly determines the efficiency, accuracy, safety, and endurance of the inspection, and is a key factor restricting the large-scale application of drones in river inspection.

[0003] Currently, extensive research has been conducted in the field of water conservancy on drone flight path planning technology. For example, Chinese patent application CN119247979A discloses an intelligent flight path planning method for reservoir inspection. This method generates full-coverage inspection flight paths for fixed structures such as reservoir dams, spillways, and reservoir shorelines, and performs basic obstacle avoidance based on terrain data, thereby improving the coverage efficiency of reservoir inspections to a certain extent. Another example is Chinese patent application CN120445226A, which discloses a drone flight path planning method for hydropower projects. This method generates refined inspection flight paths for complex three-dimensional structures based on BIM models, supports multi-angle shooting, and meets the accuracy requirements for detecting defects in engineering structures such as dams, power plants, and slopes.

[0004] However, the aforementioned existing technologies have significant limitations when applied to river patrol scenarios and cannot meet the inspection needs of long-distance linear dynamic rivers:

[0005] Regarding the technical solution disclosed in CN119247979A, firstly, the application scenarios are fundamentally mismatched. This solution is specifically designed for closed, static reservoirs and adopts a "no blind spots, full coverage" route planning logic. However, rivers are open linear networks (including main streams, tributaries, and bifurcation sections), and their core requirement is "long-distance, efficient, segmented coverage." Directly applying this solution would result in a large number of invalid flights and extremely low inspection efficiency. Secondly, it lacks a multi-task classification and priority mechanism, only supporting a single full-coverage inspection task. It fails to distinguish between different task types such as routine river patrols, sand mining special investigations, sewage outlet monitoring, and flood emergency patrols, making it impossible to perform tasks in a timely manner. In emergency situations, it interrupts routine tasks to perform emergency missions and cannot automatically increase the frequency of inspections of key areas such as sewage outlets, sand mining sites, and dangerous sections of the river. Third, it does not consider the unique environmental constraints of the river channel, completely ignoring the signal blind spots in river canyons, dense forests, and under bridges. It also cannot automatically adjust flight altitude and speed according to the river's curvature, shoreline slope, and real-time water level, making it less practical. Fourth, it lacks the ability to link with unmanned airports and coordinate with multiple drones, only supporting manual take-off and landing of a single drone. It does not integrate automatic charging of unmanned airports or cross-airport mission relay functions, and cannot overcome the single drone's endurance limit to achieve continuous inspection of long-distance river channels.

[0006] Regarding the technical solution disclosed in CN120445226A, firstly, the technical approach contradicts the needs of river channels. This solution is specifically designed for high-precision detection of point-like engineering structures, employing a low-altitude, low-speed, multi-angle shooting mode, resulting in a very small coverage area per flight, which cannot meet the rapid coverage requirements of daily patrols of rivers spanning hundreds of kilometers. Secondly, there is no dynamic route update mechanism; it adopts a pre-planned static route mode, which cannot automatically adjust the route for key inspections based on problems discovered during real-time inspections (such as sudden floating objects or temporary illegal constructions), nor can it dynamically correct the route based on real-time hydrological data, weather changes, and the real-time status of the drone (such as insufficient battery power or poor signal). Thirdly, the task constraints are singular, only considering the single constraint of structural detection accuracy, without considering time constraints (such as emergency missions requiring arrival at the scene within 30 minutes), cost constraints (such as prioritizing the use of drones with more remaining battery power), and business process constraints. Fourthly, the business integration is low, only remaining at the "route generation" stage, without deep integration with the business processes of river inspection such as task scheduling, problem identification, and closed-loop handling, and cannot optimize future routes based on historical inspection data.

[0007] In summary, existing UAV flight path planning technologies for water conservancy are mainly designed for static objects such as reservoirs and hydropower projects, and cannot adapt to the inspection needs of long-distance linear dynamic rivers. They generally suffer from problems such as a lack of multi-tasking capabilities, incomplete constraint considerations, insufficient dynamic adjustment capabilities, low levels of collaboration and automation, and poor business integration, severely restricting the large-scale and routine application of UAVs in river inspection. Therefore, there is an urgent need to develop a multi-tasking intelligent flight path planning method for UAV river patrols in water conservancy. Summary of the Invention

[0008] The purpose of this invention is to provide a multi-task intelligent route planning method for water conservancy UAV river patrol, aiming to solve the problems mentioned in the background art. The existing water conservancy UAV route planning technology is mainly designed for static objects such as reservoirs and hydropower projects, which cannot adapt to the inspection needs of long-distance linear dynamic rivers. It generally suffers from problems such as lack of multi-task capability, incomplete consideration of constraints, insufficient dynamic adjustment capability, low degree of collaboration and automation, and poor business integration.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a multi-task intelligent route planning method for water conservancy unmanned aerial vehicle (UAV) river patrol, comprising the following steps:

[0010] S1. Task Classification and Priority Ranking: Upon receiving inspection task requests, tasks are classified into three categories: daily river patrol, special inspection, and emergency patrol. The task priority is determined based on the task type, river section level, and urgency.

[0011] S2. Multi-constraint modeling: Construct a multi-constraint model that includes equipment constraints, environmental constraints, and task constraints. The equipment constraints include UAV endurance, maximum flight distance, and payload. The environmental constraints include weather conditions, terrain obstacles, and signal blind spots. The task constraints include inspection accuracy, key areas, and re-flight requirements.

[0012] S3. Initial route generation: Based on the task type and priority, and combined with the multi-constraint model, an initial inspection route is generated; among them, a fixed segmented route is generated for daily river patrol tasks, a key area densified route is generated for special inspection tasks, and a rapid response mobile route is generated for emergency patrol tasks.

[0013] S4. Route Optimization: Optimize the initial route, including multi-angle go-around optimization in key areas, range optimization, and signal blind spot avoidance optimization;

[0014] S5. Dynamic Flight Path Adjustment: During the inspection process, the drone's status, environmental changes, and problems discovered during the inspection are obtained in real time, and the flight path is dynamically adjusted.

[0015] S6. Multi-drone collaboration and task relay: When there are multiple parallel tasks or the single drone's endurance is insufficient, multi-drone task allocation and cross-drone airport task relay are carried out.

[0016] S7. Route Execution and Results Feedback: Control the UAV to perform inspection tasks according to the planned route, and feed back the inspection results and route execution data to the dispatch platform.

[0017] Preferably, in step S1, the task priority ranking rule is as follows: emergency patrol tasks have a higher priority than special inspection tasks, and special inspection tasks have a higher priority than daily river patrol tasks; among the same type of tasks, core river section tasks have a higher priority than ordinary river section tasks.

[0018] Preferably, in step S2, the environmental constraints also include real-time water level data. When the water level exceeds a preset threshold, the inspection range is automatically adjusted to avoid flooded areas.

[0019] Preferably, in step S3, the initial route generation method for daily river patrol tasks is as follows: the long river is divided into multiple segments that match the drone's endurance, and each segment generates a round-trip route parallel to the river's centerline. The route spacing is determined according to the inspection accuracy requirements.

[0020] Preferably, in step S4, the multi-angle re-flying optimization of key areas specifically involves generating multiple shooting points at different heights and angles for key areas such as sewage outlets, sand mining points, and dangerous sections of the project, to ensure comprehensive coverage.

[0021] Preferably, in step S4, the battery life optimization specifically involves: calculating the remaining battery power of the drone after completing the current route; when the remaining battery power is lower than a preset threshold, planning the nearest drone airport as the return point, or triggering a cross-airport mission relay.

[0022] Preferably, in step S5, the triggering conditions for dynamic route adjustment include:

[0023] If a suspected anomaly is detected, the flight path will be automatically adjusted to perform encrypted photography of the problem area;

[0024] If the real-time wind speed exceeds the preset safety threshold, the flight altitude will be automatically reduced or the aircraft will return to base.

[0025] If it enters a signal blind spot, it will automatically switch to a preset safe route or return to base.

[0026] Preferably, in step S6, when there are multiple parallel inspection tasks, the multi-drone task allocation adopts a distributed auction algorithm. Each drone calculates the task cost and bids based on its remaining power, current position and task priority, and selects the drone with the lowest overall cost to execute the corresponding task.

[0027] Preferably, in step S6, the cross-drone airport task relay is as follows: when a single drone cannot complete the inspection of the entire river section, task boundaries are defined between adjacent drone airports. After the first drone completes its own task segment, it returns to the airport to recharge, and the next drone takes off from its airport to continue to perform the remaining tasks.

[0028] Preferably, it also includes step S8, route self-learning optimization: based on historical inspection data and problem distribution, automatically adjust the inspection frequency and key areas of future routes, and increase the number of inspections in areas with high problem incidence.

[0029] Compared with the prior art, the beneficial effects of the present invention are:

[0030] 1. This invention divides river inspection tasks into three categories: routine river patrol, special inspection, and emergency patrol. It also establishes a scientific task priority ranking mechanism, which can automatically interrupt routine tasks to prioritize emergency tasks in emergency situations. At the same time, it automatically intensifies inspections of key areas such as sewage outlets, sand mining sites, and dangerous sections, greatly improving the efficiency and targeting of inspection tasks.

[0031] 2. This invention comprehensively considers three major categories of constraints: equipment constraints, environmental constraints, and task constraints. It specifically incorporates factors unique to rivers, such as real-time water levels, signal blind spots, and terrain obstacles. It can automatically adjust flight altitude and speed based on river curvature and shoreline slope, and automatically correct the inspection range based on real-time water level changes. This effectively avoids signal blind spots and dangerous areas, reducing the drone flight accident rate by more than 60% and ensuring the safety and reliability of inspection operations.

[0032] 3. This invention designs differentiated initial route generation methods for different task types: daily river patrols use segmented parallel routes to achieve long-distance and efficient coverage, special inspections use key area densified routes to achieve refined detection, and emergency patrols use rapid response mobile routes to achieve arrival in the shortest time.

[0033] 4. This invention can acquire the drone's status, environmental changes, and problems found during the inspection process in real time, and can automatically trigger flight path adjustments: it can perform encrypted shooting for suspected abnormal problems, automatically reduce flight altitude or return to base in bad weather, and automatically switch to safe flight paths in signal blind spots. It can flexibly respond to the dynamic changes in the river channel, ensure the integrity and effectiveness of the inspection data, and avoid inspection mission failure due to environmental changes.

[0034] 5. This invention uses a distributed auction algorithm to allocate tasks to multiple drones and supports a task relay mechanism across drone airports. It can divide a hundred-kilometer-level river into multiple sub-tasks and assign them to different drones for execution. Through task relay between adjacent airports, it can achieve long-distance continuous inspection, completely breaking through the limitation of single drone endurance and increasing the coverage distance of a single inspection to more than 3 times the original.

[0035] 6. This invention can automatically adjust the inspection frequency and key areas of future routes based on historical inspection data and problem distribution, increase the number of inspections in areas with high problem incidence, realize dynamic optimization and continuous improvement of routes, continuously improve the pertinence and effectiveness of inspections, further reduce the missed inspection rate, and provide strong support for the refined management of waterways. Attached Figure Description

[0036] Figure 1 This is an overall flowchart of the intelligent route planning method for multi-task river patrol using unmanned aerial vehicles (UAVs) according to the present invention.

[0037] Figure 2 This is a schematic diagram of the cross-unmanned airport mission relay of the present invention. Detailed Implementation

[0038] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] Please see Figure 1-2The present invention provides the following technical solution: a multi-task intelligent route planning method for water conservancy UAV river patrol, which realizes efficient and intelligent inspection of long-distance linear dynamic river channels through task classification and priority ranking, multi-constraint modeling, initial route generation, route optimization, dynamic route adjustment, multi-UAV collaboration and task relay, route execution and result feedback, and route self-learning optimization.

[0040] Furthermore, the cloud-based dispatch platform receives inspection task requests submitted by users via web or mobile devices, parses task parameters including task type, inspection river section, execution time, inspection accuracy requirements, and key areas. The system categorizes tasks into three types: routine river patrols, special inspections, and emergency patrols. Task priorities are determined based on task type, river section level, and urgency. The priority ranking rule is: emergency patrols have higher priority than special inspections, and special inspections have higher priority than routine river patrols. Within the same type of task, tasks in core river sections have higher priority than those in ordinary river sections. Core river sections include urban built-up areas, drinking water source protection areas, dangerous sections, and sections near important water conservancy projects. When multiple tasks are waiting to be executed simultaneously, the system dispatches drones to execute tasks in descending order of priority. High-priority tasks can interrupt ongoing low-priority tasks; once the high-priority task is completed, the low-priority task can resume execution. For example, when the system receives a routine river patrol task for the upper reaches of the Tongshi River and an emergency flood patrol task for the Yongfa Town section of the Nandu River at the same time, the system immediately suspends the routine river patrol task and dispatches the drone closest to the Yongfa Town section of the Nandu River to carry out the emergency patrol task. After the emergency task is completed, the routine river patrol task for the upper reaches of the Tongshi River will be resumed.

[0041] Furthermore, based on this, the system constructs a multi-constraint model, including equipment constraints, environmental constraints, and task constraints, as the foundation for flight path planning. Equipment constraints include a maximum drone endurance of 40 minutes, a maximum flight distance of 20 kilometers, a maximum payload of 1 kg, a maximum flight speed of 20 m / s, and a wind resistance of level 6. The system acquires the remaining battery power, current location, and load status of each drone in real time, calculating the maximum range of its executable tasks. Environmental constraints include weather conditions, terrain obstacles, signal blind spots, and real-time water level data. Weather conditions are obtained through access to the national meteorological API; drones are prohibited from taking off when wind speeds exceed level 6 or rainfall exceeds 5 mm / h. Terrain obstacle data comes from digital elevation models and vector maps, including bridges, high-voltage lines, buildings, and mountains. Signal blind spot data is obtained through prior inspections and user annotations, marking areas as prohibited or restricted flight zones. Real-time water level data comes from riverside hydrological stations; when water levels exceed preset warning levels, the system automatically adjusts the inspection range to avoid flooded shoreline areas. Task constraints include inspection accuracy, key areas, and go-around requirements. Routine river patrol tasks require a ground resolution of ≤5cm / pixel, while special inspection tasks require a ground resolution of ≤2cm / pixel. Key areas include sewage outlets, sand mining sites, dangerous sections of dikes, and construction areas, requiring multi-angle re-flying. The number of re-flying operations is determined according to the task requirements, with one re-flying operation for general areas and more than two re-flying operations for key areas.

[0042] Furthermore, based on task type and priority, and combined with a multi-constraint model, the system uses an improved A* algorithm to generate initial inspection routes. For routine river patrols, fixed segmented routes are generated: First, long river sections are divided into multiple segments according to the drone's endurance, each segment not exceeding 15 kilometers in length, ensuring a single drone can complete the inspection of one segment. Then, within each segment, a round-trip route parallel to the river's centerline is generated. The route spacing is determined based on inspection accuracy requirements: 30 meters for routine patrols, 20 meters horizontally from the shoreline, and a flight altitude of 50 meters. For specialized inspection tasks, encrypted routes are generated for key areas: Based on the regular routes, a circling encrypted route is generated for key areas. The drone's flight altitude is reduced to 20 meters, generating four different angle shooting points around the key area: directly in front, to the right, directly behind, and to the left. Each waypoint is paused for 2 seconds to capture images, ensuring comprehensive coverage of the key area. Emergency patrol missions generate rapid response mobile routes: with the goal of reaching the scene as quickly as possible, the shortest direct flight path from the nearest unmanned airport to the incident site is planned, avoiding terrain obstacles and no-fly zones; after arriving at the incident site, a spiral search route is generated, gradually expanding outward from the center to cover the area affected by the incident, with a flight altitude of 30 meters and a flight speed of 5 m / s, and real-time transmission of high-definition video footage.

[0043] Furthermore, the system performs multi-objective optimization on the generated initial flight path, with optimization objectives including shortest flight distance, minimum flight time, highest coverage efficiency, and lowest energy consumption. During multi-angle go-around optimization in key areas, the system automatically calculates the optimal shooting angle and altitude, generating the optimal sequence of shooting waypoints to avoid duplicate or missed shots. For example, for areas with damaged riverbanks, the system generates shooting routes perpendicular to the riverbank, shooting the front and sides of the damaged area from different heights to ensure a clear display of the extent of damage. During endurance optimization, the system calculates in real time the battery power required for the drone to complete the current flight path, reserving 20% ​​for return. When the calculated remaining battery power is insufficient to complete the current flight path, the system automatically plans the nearest drone airport as the return point, or triggers a cross-airport task relay mechanism to assign the remaining tasks to drones at adjacent airports. During signal blind spot avoidance optimization, the system automatically avoids pre-marked signal blind spots during flight path planning. If avoidance is impossible due to the river's course, a hovering point is set at the edge of the signal blind spot. After the drone reaches the hovering point, it first checks the signal strength; if the signal strength meets the requirements, it continues flying; otherwise, it returns along the original route to the signal recovery area.

[0044] Furthermore, during the drone's inspection mission, the system acquires real-time data on the drone's status, environmental changes, and identified issues, dynamically adjusting the flight path based on the actual situation. The triggering conditions for dynamic flight path adjustments fall into three categories: First, the detection of suspected anomalies. When the AI ​​recognition model detects a suspected anomaly in real time, the system automatically adjusts the flight path, controlling the drone to fly to the problem area, lowering its flight altitude to 20 meters, and increasing the shooting frequency to 60 frames per second to encrypt the image of the problem area and obtain clear, detailed images. Second, real-time wind speed exceeding a preset safety threshold. When the weather station detects a real-time wind speed exceeding level 6, the system immediately controls the drone to lower its flight altitude to below 30 meters. If the wind speed continues to increase to level 8, the system controls the drone to return to the nearest drone airport. Third, entering a signal blind zone. When the drone enters a signal blind zone and the signal strength is below -100dBm, the system automatically switches to a preset safe flight path, controlling the drone to return along the original route to the signal recovery area, while simultaneously sending a warning message to management personnel.

[0045] Furthermore, when multiple parallel inspection tasks exist or a single drone's battery life is insufficient, the system performs multi-drone task allocation and cross-drone airport task relay. Multi-drone task allocation employs a distributed auction algorithm: each drone calculates its task cost based on its remaining battery power, current location, and task priority. The task cost calculation formula comprehensively considers three factors: flight distance, remaining battery power, and task priority. Each drone broadcasts its calculated task cost through the cloud scheduling platform, and the system selects the drone with the lowest overall cost to execute the corresponding task. During cross-drone airport task relay, when a single drone cannot complete the inspection task for the entire river section, the system defines a task boundary between adjacent drone airports. The task boundary is set in the middle of the two drone airports, ensuring that after the preceding drone takes off from airport A and completes task segment A, it has enough remaining battery power to return to airport A; and after the following drone takes off from airport B and completes task segment B, it has enough remaining battery power to return to airport B. After completing its task segment, the preceding drone automatically returns to airport A to recharge; the following drone simultaneously takes off from airport B to continue executing the remaining task segment B, achieving continuous inspection of long-distance river channels.

[0046] Furthermore, the cloud-based dispatch platform distributes the planned flight routes to the corresponding unmanned aerial vehicle (UAV) airports, which then control the UAVs to automatically execute inspection tasks according to the routes. During the execution of the routes, the UAVs transmit flight status data and image data back to the cloud-based dispatch platform in real time. After completing the inspection task, the UAVs automatically return to the UAV airport to recharge, and the system automatically generates an inspection report, including inspection time, inspected river section, flight mileage, flight time, a list of problems found, on-site photos and videos, and other information.

[0047] Furthermore, the system establishes a self-learning mechanism for its flight routes, continuously optimizing future inspection routes based on historical inspection data and problem distribution. The system statistically analyzes the frequency, type, and location of problems in each river segment over the past six months, generating a heatmap of river problems. For high-problem areas, the system automatically marks them as key areas of concern, increasing inspection frequency from once a week to twice a week; for low-problem areas, the system appropriately reduces inspection frequency from once a week to once every two weeks. Simultaneously, the system adjusts the key areas and shooting angles of the flight routes, optimizes route spacing and flight altitude, and enhances the targeting and effectiveness of inspections based on problem distribution.

[0048] Working principle: Users submit inspection task requests through a cloud-based scheduling platform. The system categorizes tasks and determines priorities. Next, the system constructs a multi-constraint model of equipment, environment, and task, and uses an improved A* algorithm to generate differentiated initial inspection routes based on task type. The initial routes are then optimized for key areas, extended flight time, and signal blind spot avoidance. During the drone's route execution, the system monitors the drone's status, environmental changes, and inspection results in real time, dynamically adjusting the route based on trigger conditions. When multiple parallel tasks exist or a single drone's flight time is insufficient, the system uses a distributed auction algorithm to allocate tasks across multiple drones and supports task relay across drone airports. After completing the inspection task, the drone feeds back the inspection results and route execution data to the scheduling platform. Finally, the system learns from historical inspection data and problem distribution, automatically adjusting the inspection frequency and key areas of future routes to achieve continuous route optimization and intelligent, efficient river inspection.

[0049] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-task intelligent route planning method for river patrol by unmanned aerial vehicles (UAVs), characterized in that, Includes the following steps: S1. Task Classification and Priority Ranking: Upon receiving inspection task requests, tasks are classified into three categories: daily river patrol, special inspection, and emergency patrol. The task priority is determined based on the task type, river section level, and urgency. S2. Multi-constraint modeling: Construct a multi-constraint model that includes equipment constraints, environmental constraints, and task constraints. The equipment constraints include UAV endurance, maximum flight distance, and payload. The environmental constraints include weather conditions, terrain obstacles, and signal blind spots. The task constraints include inspection accuracy, key areas, and re-flight requirements. S3. Initial route generation: Based on the task type and priority, and combined with the multi-constraint model, an initial inspection route is generated; among them, a fixed segmented route is generated for daily river patrol tasks, a key area densified route is generated for special inspection tasks, and a rapid response mobile route is generated for emergency patrol tasks. S4. Route Optimization: Optimize the initial route, including multi-angle go-around optimization in key areas, range optimization, and signal blind spot avoidance optimization; S5. Dynamic Flight Path Adjustment: During the inspection process, the drone's status, environmental changes, and problems discovered during the inspection are obtained in real time, and the flight path is dynamically adjusted. S6. Multi-drone collaboration and task relay: When there are multiple parallel tasks or the single drone's endurance is insufficient, multi-drone task allocation and cross-drone airport task relay are carried out. S7. Route Execution and Results Feedback: Control the UAV to perform inspection tasks according to the planned route, and feed back the inspection results and route execution data to the dispatch platform.

2. The method for multi-task intelligent route planning of unmanned aerial vehicle (UAV) river patrol according to claim 1, characterized in that, In step S1, the priority ranking rule for tasks is as follows: emergency patrol tasks have a higher priority than special inspection tasks, and special inspection tasks have a higher priority than daily river patrol tasks; among tasks of the same type, core river section tasks have a higher priority than ordinary river section tasks.

3. The method for multi-task intelligent route planning of unmanned aerial vehicle (UAV) river patrol according to claim 1, characterized in that, In step S2, the environmental constraints also include real-time water level data. When the water level exceeds a preset threshold, the inspection range is automatically adjusted to avoid flooded areas.

4. The method for multi-task intelligent route planning of unmanned aerial vehicle (UAV) river patrol according to claim 1, characterized in that, In step S3, the initial route generation method for daily river patrol tasks is as follows: the long river channel is divided into multiple segments that match the drone's endurance, and each segment generates a round-trip route parallel to the center line of the river channel. The route spacing is determined according to the inspection accuracy requirements.

5. The method for multi-task intelligent route planning of unmanned aerial vehicle (UAV) river patrol according to claim 1, characterized in that, In step S4, the multi-angle re-flight optimization of key areas specifically involves generating multiple shooting points at different heights and angles for key areas such as sewage outlets, sand mining points, and dangerous sections of the project, to ensure comprehensive coverage.

6. The method for multi-task intelligent route planning of unmanned aerial vehicle (UAV) river patrol according to claim 1, characterized in that, In step S4, the battery life optimization specifically involves: calculating the remaining battery power of the drone after completing the current route; when the remaining battery power is lower than a preset threshold, planning the nearest drone airport as the return point, or triggering a cross-airport mission relay.

7. The method for multi-task intelligent route planning of unmanned aerial vehicle (UAV) river patrol according to claim 1, characterized in that, In step S5, the triggering conditions for dynamic route adjustment include: If a suspected anomaly is detected, the flight path will be automatically adjusted to perform encrypted photography of the problem area; If the real-time wind speed exceeds the preset safety threshold, the flight altitude will be automatically reduced or the aircraft will return to base. If it enters a signal blind spot, it will automatically switch to a preset safe route or return to base.

8. The method for multi-task intelligent route planning of unmanned aerial vehicle (UAV) river patrol according to claim 1, characterized in that, In step S6, when there are multiple parallel inspection tasks, the multi-drone task allocation adopts a distributed auction algorithm. Each drone calculates the task cost and bids based on its remaining power, current position and task priority, and selects the drone with the lowest overall cost to execute the corresponding task.

9. A multi-task intelligent route planning method for river patrol by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, In step S6, the relay of tasks across unmanned airports is as follows: when a single drone cannot complete the inspection of the entire river section, task boundaries are defined between adjacent unmanned airports. After the first drone completes its own task segment, it returns to the airport to recharge, and the second drone takes off from its airport to continue to perform the remaining tasks.

10. A multi-task intelligent route planning method for river patrol by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, It also includes step S8, route self-learning optimization: based on historical inspection data and problem distribution, automatically adjust the inspection frequency and key areas of future routes, and increase the number of inspections in areas with high problem incidence.

Images