Dynamic planning method and system for patrol path of unmanned aerial vehicle in three-dimensional scene

Abstract

The invention discloses a dynamic planning method and system for a patrol path of an unmanned aerial vehicle in a three-dimensional scene. The method comprises the following steps: S1, acquiring an initial flight path according to a topographic map; s2, flying according to the initial flight path, scanning flight area depth information in real time at each flight path point in the initial flight path, and obtaining a distance measurement value in front of the unmanned aerial vehicle and a depth actual measurement value below the unmanned aerial vehicle; s3, comparing the depth measured value with a depth reference value in the initial flight path, and if an absolute value of a difference value between a plurality of continuous depth measured values and the depth reference value H is greater than a preset first threshold value, determining that the terrain is changed, enabling the unmanned aerial vehicle to enter a slow-down mode, and reducing the flight speed; s4, according to the distance measurement value in front of the unmanned aerial vehicle and the depth actual measurement value below the unmanned aerial vehicle, whether a collision risk exists or not is judged, and automatic obstacle avoidance and dynamic route planning are carried out; and if there is a collision risk, an image in front of the unmanned aerial vehicle is collected, and image processing is utilized to make an obstacle avoidance decision. According to the invention, automatic obstacle avoidance and inspection are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and in particular to a dynamic planning method for an unmanned aerial vehicle inspection path in a three-dimensional scene. Background Art

[0002] With the rapid development of aviation science and technology, drone technology has been rapidly improved. In recent years, the use of drones to replace manual operations has become a hot topic. In particular, the use of drones for inspections can accurately and comprehensively check the operation of the line, greatly improving the efficiency and quality of inspections.

[0003] Currently, drones are used to patrol the destination when natural disasters occur. However, drones are generally controlled manually by people, which is difficult and prone to collision accidents. Alternatively, the original map is used for path planning. However, after a natural disaster occurs, the topography of the disaster site may change, so obstacles are likely to appear on the planned path, making it difficult for drones to reach the target location to perform patrol tasks. Summary of the invention

[0004] The purpose of the present invention is to provide a dynamic planning method and system for UAV inspection paths in a three-dimensional scene, which can effectively respond to geological changes in disaster sites, automatically perform obstacle avoidance and inspections, and ensure safe and accurate execution of inspection tasks.

[0005] In order to achieve one of the above purposes, the present invention adopts the following technical solution:

[0006] A dynamic planning method for a UAV inspection path in a three-dimensional scene includes the following steps:

[0007] S1: Obtain the initial flight path based on the existing terrain map;

[0008] S2: Fly according to the initial flight path in step S1, and at each flight path point W in the initial flight path i Scan the depth information of the flight area in real time to obtain the distance measurement value S(W) in front of the drone i ) and the measured depth h(W i );

[0009] S3: The measured depth value h(W i ) and the depth reference value H(W) in the initial flight path i ) for comparison, if there are two or more consecutive depth measured values ​​h(W i ) and depth reference value H(W i ) when the absolute value of the difference is greater than a preset first threshold, the terrain has changed, and the UAV enters a slow-flying mode and reduces the flight speed;

[0010] S4: According to the distance measurement value S(W i ) and the measured depth h(W i ), determine whether there is a collision risk, perform automatic obstacle avoidance and dynamic planning of the route; if there is a collision risk, collect the image in front of the drone, and use the image processing to make obstacle avoidance decisions, specifically including:

[0011] First, obtain the image in front of the drone, and calibrate the ranging point O in the image so that the ranging point O is located at the exact center of the image;

[0012] Then, the distance between the drone position and the distance measurement point O is the distance measurement value S(W i ), if the fixed viewing angle of the camera is α, the sum of the angles formed by the ranging line and the viewing angle edge lines on both sides horizontally is α, that is, the sum of the angles formed by the ranging line and the viewing angle edge lines on both sides vertically is α';

[0013] In the image, the sum of the distances from the distance measuring point O to the edges of the left and right viewing angles is L1, and the sum of the distances from the distance measuring point O to the edges of the upper and lower viewing angles is L1';

[0014] Extract the obstacle boundary and determine the distances between the ranging point O and the left, right, upper and lower boundaries as L2, L3, L2' and L3' respectively. At this time, the left angle β, right angle γ, upper angle β' and lower angle γ' between the ranging line and the boundary line are obtained;

[0015]

[0016]

[0017] The actual distances S(L2), S(L3), S(L2'), S(L3') corresponding to L2, L3, L2', and L3' are calculated by trigonometric function. S(L2) = S(W i )×tanβ, S(L3)=S(W i )×tanγ,S(L2')=S(W i )×tanβ', S(L3')=S(W i )×tanγ';

[0018] The obstacle avoidance direction is determined by selecting the shortest boundary distance between the current position of the drone and the upper, lower, left, and right boundaries of the obstacle in front. At the same time, the obstacle avoidance flight distance is determined based on the actual distance corresponding to the shortest boundary distance plus the obstacle avoidance safety distance threshold S.

[0019] Furthermore, obtaining the image in front of the drone includes: fusing distance information obtained by a ranging sensor of the drone with data collected by a camera of the drone to obtain the image in front of the drone.

[0020] Furthermore, the determination of whether there is a collision risk, automatic obstacle avoidance and dynamic planning of the route specifically include:

[0021] a. If the distance measurement value S(W i ) is greater than the preset first collision threshold, and the measured depth value h(W i ) is greater than the preset second collision threshold, it is determined that the UAV does not have a collision risk, and the UAV flies along the original flight path to the target point;

[0022] b. If the distance measurement value S(W i ) is greater than the preset first collision threshold, and the measured depth value h(W i ) is less than or equal to the preset second collision threshold, the drone is judged to have a bottom collision risk. At this time, the drone stops moving forward and rises to the depth measured value h(W below the drone). i ) is greater than the preset second collision threshold, continue to fly to the previous target point;

[0023] c. If the distance measurement value S(W i ) is less than or equal to the preset first collision threshold, and the measured depth value h(W i ) is greater than a preset second collision threshold, it is determined that the UAV has a front collision risk. At this time, the image in front of the UAV is collected and image processing is used to make obstacle avoidance decisions;

[0024] d. If the distance measurement value S(W i ) is less than or equal to the preset first collision threshold, and the measured depth value h(W i ) is less than or equal to the preset second collision threshold, it is determined that the drone has a risk of collision with the bottom and the front. At this time, the drone stops moving forward and rises to the depth measured value h(W below the drone). i ) is greater than the preset second collision threshold, the image in front of the drone is collected and image processing is used to make obstacle avoidance decisions.

[0025] Furthermore, the S1: obtaining an initial flight path according to an existing topographic map; specifically includes:

[0026] S101: Perform three-dimensional environment modeling according to the existing topographic map to obtain a three-dimensional map;

[0027] S102: Determine a plurality of flight path points W from the patrol area of ​​the three-dimensional map i, i represents the number of path points, and the initial flight path is obtained;

[0028] S103: Mapping the initial flight path to the three-dimensional map and obtaining each flight path point W i The depth reference value H(W i ).

[0029] Further, the S102: determining a flight path point W from the patrol area of ​​the three-dimensional map i , get the initial flight path; including:

[0030] First, create a preliminary flight path from a 3D map based on the patrol area;

[0031] Secondly, setting constraints and constructing an objective function, wherein the constraints include at least: a flight path cost constraint, an obstacle avoidance cost constraint and a flight altitude cost constraint, and the objective function includes: the shortest flight path and the smallest flight threat;

[0032] Finally, multiple UAV path points W of the preliminary flight path are determined according to the constraint conditions and the objective function. i Position and get the initial flight path.

[0033] Furthermore, the constraints include at least: a flight path cost constraint, an obstacle avoidance cost constraint, and a flight altitude cost constraint, and the objective function includes: the shortest flight path and the smallest flight threat, specifically including:

[0034] The flight path cost constraint:

[0035] The flight path includes a plurality of flight path points W i , flight path point W i The coordinates are W i (x i ,y i , z i ), the distance between two adjacent flight path points is D(W i , W i+1 );

[0036] The total flight path cost is

[0037] The obstacle avoidance cost constraint:

[0038] Obtain the contour map of the inspection area.

[0039] Assume that the measured value of the drone's flight altitude depth is h(W i ), determine the shortest distance s between the UAV and the contour line min, the corresponding contour lines are determined from the contour topographic map, and the width d outside the contour lines is set as the regional threat area;

[0040] For any obstacle, the UAV is at a flight height of h(W i ), its risk cost R j It is expressed as:

[0041]

[0042] Among them, θ is the set first cost coefficient, j is the number of obstacle avoidance points during flight;

[0043] The total risk cost is

[0044] The flight altitude cost constraint:

[0045] Set threat zone: above minimum altitude h min , below the minimum height h min + a height range of one unit height h0 (h min ,h min +h0), h0 is determined based on the height of local vegetation, and the minimum height h min Determined by terrain height in a three-dimensional map;

[0046] Flight Waypoint W i The flight altitude cost h i for:

[0047]

[0048] Among them, ɡ is the set second cost coefficient; the maximum height h max The maximum flight altitude of the drone;

[0049] The total height cost is

[0050] The flight path is the shortest f1(A):

[0051] The minimum flight threat is f2(A):

[0052] Furthermore, the multiple UAV path points W of the preliminary flight path are determined according to the constraint conditions and the objective function. i Position, get the initial flight path, including:

[0053] Substitute the constraints and objective function into the MATLAB mathematical tool and use the fmincon function to solve and calculate the multiple UAV path points W of the preliminary flight path. iPosition, a preliminary flight path can be obtained.

[0054] Furthermore, the flight speed is reduced to 20%-50%.

[0055] In order to achieve the second of the above objectives, the present invention adopts the following technical solution:

[0056] A dynamic planning system for an unmanned aerial vehicle inspection path in a three-dimensional scene comprises an unmanned aerial vehicle and a flight control device, wherein the flight control device adopts any of the above-mentioned dynamic planning methods for an unmanned aerial vehicle inspection path in a three-dimensional scene.

[0057] In order to achieve the third of the above objectives, the present invention adopts the following technical solutions:

[0058] A flight control device comprises: a processor and a memory storing instructions executable by the processor, wherein the processor executes any one of the above-mentioned methods for dynamically planning a patrol path of an unmanned aerial vehicle in a three-dimensional scene.

[0059] Beneficial effects of the present invention:

[0060] The method of the present invention first uses three-dimensional maps, contour maps, etc. to construct constraints and objective functions to obtain an initial flight path, and then performs a flight mission according to the initial flight path. During the flight, the drone obtains depth reference information based on the initial flight path to determine whether the terrain has changed. If so, the flight speed is slowed down to avoid collisions caused by the drone flying too fast. At the same time, the drone performs automatic obstacle avoidance during the flight, and combines image recognition to determine the best obstacle avoidance direction. After the obstacle is successfully avoided, it returns to the original flight path to continue to perform the flight inspection task. Therefore, the present invention can effectively target natural disasters, automatically perform obstacle avoidance and inspections, ensure the safe and accurate execution of inspection tasks, and has a high practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a flow chart of a specific embodiment 1 of the present invention;

[0062] Figure 2 A contour map of the inspection area of ​​the specific embodiment 1 of the present invention;

[0063] Figure 3 It is a schematic diagram of calibration of the ranging point O in the image in the specific embodiment 1 of the present invention;

[0064] Figure 4 for Figure 3 Schematic diagram of the horizontal position of the middle ranging point O and the drone;

[0065] Figure 5 for Figure 3 Schematic diagram of the vertical position of the mid-range point O and the UAV. DETAILED DESCRIPTION

[0066] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Specific embodiment 1:

[0068] See also Figures 1 to 5 As shown, a dynamic planning method for a drone inspection path in a three-dimensional scene is characterized by comprising the following steps:

[0069] S1: Obtain the initial flight path based on the existing terrain map;

[0070] include:

[0071] S101: Performing three-dimensional environment modeling according to an existing topographic map to obtain a three-dimensional map. The existing topographic map may be a direct three-dimensional map, that is, the three-dimensional map has height information; the existing topographic map may also be a two-dimensional map, and modeling is performed by a two-dimensional map + height information method to obtain a three-dimensional map;

[0072] S102: Determine a plurality of flight path points W from the patrol area of ​​the three-dimensional map i , i represents the number of path points, and the initial flight path is obtained;

[0073] Specifically include:

[0074] First, a preliminary flight path is created from the three-dimensional map based on the patrol area. The preliminary flight path consists of multiple drone path points W i The construction process of this preliminary flight path is as follows:

[0075] Mark the take-off point S and patrol point IP from the 3D map i And a return point R, the take-off point S and the return point R may be the same.

[0076] Take off the drone from the take-off point S to the inspection point IP i There are multiple UAV path points between them. This scheme obtains the positions of multiple UAV path points by configuring multiple constraints and objective functions;

[0077] Secondly, setting constraints and constructing an objective function, wherein the constraints include at least: a flight path cost constraint, an obstacle avoidance cost constraint and a flight altitude cost constraint, and the objective function includes: the shortest flight path and the smallest flight threat;

[0078] The constraints include at least: flight path cost constraint, obstacle avoidance cost constraint and flight altitude cost constraint, and the objective function includes: shortest flight path and minimum flight threat, specifically including:

[0079] The flight path cost constraint:

[0080] Assume that the flight path point A includes multiple flight path points with coordinates W i (x i ,y i , z i ), the distance between two adjacent flight path points is D(W i , W i+1 );

[0081] The total flight path cost is

[0082] The obstacle avoidance cost constraint:

[0083] Because the drone will bring additional flight energy consumption, flight distance and flight time when avoiding obstacles. Therefore, the total obstacle avoidance cost in the flight path is calculated through the obstacle avoidance cost constraint; it is mainly used in regional inspections, so the main obstacles are mountains, hills, buildings, etc. The characteristics of mountains and hills are that as the height increases, the obstacle boundary will shrink, thereby reducing the obstacle area. The characteristic of buildings is that as the height increases, the obstacle boundary remains unchanged.

[0084] Obtain the contour map of the inspection area.

[0085] Assume that the UAV is at a flight altitude h(W i ), the corresponding contour line is determined from the contour line topographic map, and then the shortest distance s between the UAV and the contour line is determined min , set the width d outside the contour line as the regional threat area, in this specific embodiment: d is 5-20m;

[0086] For any obstacle, the UAV is at a flight height of h(W i ), its risk cost R j It is expressed as:

[0087]

[0088] Among them, θ is the set first cost coefficient, θ is 0.5, j is the number of obstacle avoidance points during flight; max indicates that there is a safety problem;

[0089] The total risk cost is

[0090] The flight altitude cost constraint:

[0091] The drone cannot exceed the set maximum altitude during flight. max , nor can it be lower than the set minimum height h min , set the threat zone: above the minimum height h min , and minimum height h min + a height range of one unit height h0 (h min ,h min +h0), h0 is determined based on the height of local vegetation, and the minimum height h min Determined by the terrain height in the 3D map, the highest height h max The maximum flight altitude of the drone;

[0092] Flight Waypoint W i The flight altitude cost h i for:

[0093]

[0094] Among them, ɡ is the set second cost coefficient, ɡ is 0.5; max means there is a safety problem;

[0095] The total height cost is

[0096] The flight path is the shortest f1(A):

[0097] The minimum flight threat is f2(A):

[0098] n and m are natural numbers.

[0099] Finally, multiple UAV path points W of the preliminary flight path are determined according to the constraint conditions and the objective function. i Position and get the initial flight path.

[0100] Substitute the above constraints and objective function into the mat lab mathematical tool, and use the fmincon function to solve and calculate the multiple UAV path points W of the preliminary flight path. i The preliminary flight path can be obtained by

[0101] S103: Mapping the initial flight path to the three-dimensional map and obtaining each flight path point W i The depth reference value H(W i ).

[0102] S2: Fly according to the initial flight path in step S1, and use the ranging sensor to scan each flight path point W in the initial flight path. iScan the depth information of the flight area in real time to obtain the distance measurement value S(W) in front of the drone i ) and the measured depth h(W i );

[0103] S3: The measured depth value h(W i ) and the depth reference value H(W) in the initial flight path i ) for comparison, if there are multiple continuous depth measured values ​​h(W i ) and depth reference value H(W i )When the absolute value of the difference is greater than a preset first threshold value, the preset first threshold value is between 5-20m, then there is a change in the terrain, and the UAV enters a slow-motion mode and reduces the flight speed, wherein the reduced flight speed is reduced to 20%-50% of the existing flight speed.

[0104] S4: According to the distance measurement value S(W i ) and the measured depth h(W i ), determine whether there is a collision risk, perform automatic obstacle avoidance and dynamic route planning; if there is a collision risk, collect the image in front of the drone, and use the image processing to make obstacle avoidance decisions. Specifically include:

[0105] a. If the distance measurement value S(W i ) is greater than the preset first collision threshold, and the measured depth value h(W i ) is greater than the preset second collision threshold, it is judged that there is no collision risk for the UAV, and the UAV flies to the target point along the original flight path; the second collision threshold and the second collision threshold are between 5-20m,

[0106] b. If the distance measurement value S(W i ) is greater than the preset first collision threshold, and the measured depth value h(W i ) is less than or equal to the preset second collision threshold, the drone is judged to have a bottom collision risk. At this time, the drone stops moving forward and rises to the depth measured value h(W below the drone). i ) is greater than the preset second collision threshold, continue to fly to the previous target point;

[0107] c. If the distance measurement value S(W i ) is less than or equal to the preset first collision threshold, and the measured depth value h(W i ) is greater than a preset second collision threshold, it is determined that the UAV has a front collision risk. At this time, the image in front of the UAV is collected and image processing is used to make obstacle avoidance decisions;

[0108] Specifically include:

[0109] First, obtain the image in front of the drone, fuse the distance information obtained by the drone's ranging sensor with the data collected by the drone's camera to obtain the image in front of the drone, and calibrate the ranging point O in the fused image so that the ranging point O is located in the center of the image; the ranging point O is a point located in front of the drone, which is captured by the camera. If there are multiple objects in the image, each object uses the same calibration method.

[0110] Then, the distance between the drone position and the distance measurement point O is the distance measurement value S(W i ),like Figure 3 , Figure 4 and Figure 5 As shown, if the fixed viewing angle of the camera is α, if the obstacle is regarded as a plane, the sum of the angles formed by the ranging line and the viewing angle edge lines on both sides horizontally is α, that is, the sum of the angles formed by the ranging line and the viewing angle edge lines on both sides vertically is α';

[0111] In the image, the sum of the distances from the ranging point O to the edges of the left and right viewing angles is L1, which is determined by counting pixels. The sum of the distances from the ranging point O to the edges of the upper and lower viewing angles is L1'.

[0112] Extract the obstacle boundary and determine the distances between the ranging point O and the left, right, upper and lower boundaries as L2, L3, L2' and L3' respectively. At this time, the left angle β, right angle γ, upper angle β' and lower angle γ' between the ranging line and the boundary line are obtained;

[0113]

[0114]

[0115] α' calculates the actual distances S(L2), S(L3), S(L2'), S(L3') corresponding to L2, L3, L2', L3' through trigonometric functions, S(L2) = S(W i )×tanβ, S(L3)=S(W i )×tanγ,S(L2')=S(W i )×tanβ', S(L3')=S(W i )×tanγ';

[0116] The above formula is used to calculate the distance between the current position of the drone and the upper, lower, left and right boundaries of the obstacle in front. The shortest boundary distance is selected to determine the obstacle avoidance direction. At the same time, the obstacle avoidance flight distance is determined based on the actual distance plus the obstacle avoidance safety distance threshold S. The obstacle avoidance safety distance threshold S is the first collision threshold or the second collision threshold.

[0117] For example, if the upper direction is selected as the obstacle avoidance direction, the actual distance between the drone and the upper boundary is S(L2'), and the obstacle avoidance flight distance is S(L2')+S, where S is the obstacle avoidance distance threshold. At this time, first ascend S(L2')+S, then fly forward, and after crossing the obstacle, fly to the target point.

[0118] d. If the distance measurement value S(W i ) is less than or equal to the preset first collision threshold, and the measured depth value h(W i ) is less than or equal to the preset second collision threshold, it is determined that the drone has a risk of collision with the bottom and the front. At this time, the drone stops moving forward and rises to the depth measured value h(W below the drone). i ) is greater than the preset second collision threshold, the image in front of the drone is collected again, and image processing is used to make obstacle avoidance decisions, which is the same as the above obstacle avoidance decisions.

[0119] In this specific embodiment, the drone flies according to the initial flight path and scans the depth information of the area in real time, thereby performing obstacle avoidance and dynamic route planning to get from point A to IP. i Take point as an example;

[0120] Take off from point A and fly along the initial path. At each flight path point W i Use the rangefinder to scan and get the range value S(W) in front of the drone i ) and the measured depth h(W i );

[0121] The measured depth value h(W i ) and depth reference value H(W i ) for comparison;

[0122] When there are two or more consecutive depth measured values ​​h(W i ) and depth reference value H(W i ) is greater than the preset first threshold, indicating that the terrain has changed, the drone enters the slow-motion mode, and the flight speed is reduced by 20%-50%. At the same time, the distance measurement value S(W) in front of the drone is continuously detected. i ) and the measured depth h(W i ) for automatic obstacle avoidance.

[0123] After the obstacle avoidance is completed at the obstacle point, the drone flies in a straight line to the target point. If an obstacle is encountered again, the drone avoids the obstacle again according to the above steps S2 to S4. At the same time, the drone continues to fly until it returns to the initial flight path to complete the inspection task. Specific embodiment 2:

[0125] This specific embodiment provides a dynamic planning system for UAV patrol paths in three-dimensional scenes, including a UAV and a flight control device. The flight control device adopts a dynamic planning method for UAV patrol paths in three-dimensional scenes described in specific embodiment 1. This specific embodiment will not repeat the dynamic planning method in specific embodiment 1. Specific embodiment 3:

[0127] A flight control device includes: a processor and a memory storing executable instructions of the processor, wherein the processor executes a dynamic planning method for an unmanned aerial vehicle inspection path in a three-dimensional scene described in specific embodiment 1. The dynamic planning method in specific embodiment 1 will not be repeated in this specific embodiment.

[0128] The technical solution provided by the present invention is described in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be pointed out that for ordinary technicians in this technical field, without departing from the principle of the present invention, the present invention can also be improved and modified in a number of ways, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

Claims

1. A dynamic planning method for a drone inspection path in a three-dimensional scene, characterized in that: The steps include: S1: Obtain the initial flight path based on the terrain map; S2: Fly according to the initial flight path in step S1, and at each flight path point W in the initial flight path i Scan the depth information of the flight area in real time to obtain the distance measurement value S(W) in front of the drone i ) and the measured depth h(W i ); S3: The measured depth value h(W i ) and the depth reference value H(W) in the initial flight path i ) for comparison, if there are two or more consecutive depth measured values ​​h(W i ) and depth reference value H(W i ) when the absolute value of the difference is greater than a preset first threshold, the terrain has changed, and the UAV enters a slow-motion mode and reduces the flight speed; S4: According to the distance measurement value S(W i ) and the measured depth h(W i ), determine whether there is a collision risk, and perform automatic obstacle avoidance and dynamic route planning; If there is a risk of collision, the image in front of the drone is collected and the image processing is used to make obstacle avoidance decisions, including: First, obtain the image in front of the drone, and calibrate the ranging point O in the image so that the ranging point O is located at the exact center of the image; Then, the distance between the drone position and the distance measurement point O is the distance measurement value S(W i ), if the fixed viewing angle of the camera is α, the sum of the angles formed by the ranging line and the viewing angle edge lines on both sides horizontally is α, that is, the sum of the angles formed by the ranging line and the viewing angle edge lines on both sides vertically is α'; In the image, the sum of the distances from the distance measuring point O to the edges of the left and right viewing angles is L1, and the sum of the distances from the distance measuring point O to the edges of the upper and lower viewing angles is L1'; Extract the obstacle boundary and determine the distances between the ranging point O and the left, right, upper and lower boundaries as L2, L3, L2' and L3' respectively. At this time, the left angle β, right angle γ, upper angle β' and lower angle γ' between the ranging line and the boundary line are obtained; The actual distances S(L2), S(L3), S(L2'), S(L3') corresponding to L2, L3, L2', and L3' are calculated by trigonometric function. S(L2) = S(W i )×tanβ, S(L3)=S(W i )×tanγ,S(L2')=S(W i )×tanβ', S(L3')=S(W i )×tanγ'; The obstacle avoidance direction is determined by selecting the shortest boundary distance between the current position of the drone and the upper, lower, left, and right boundaries of the obstacle in front. At the same time, the obstacle avoidance flight distance is determined based on the actual distance corresponding to the shortest boundary distance plus the obstacle avoidance safety distance threshold S.

2. According to the method for dynamic planning of a drone inspection path in a three-dimensional scene in claim 1, it is characterized by: The obtaining of the image in front of the drone includes: fusing distance information obtained by a ranging sensor of the drone with data collected by a camera of the drone to obtain the image in front of the drone.

3. The method for dynamically planning a UAV inspection path in a three-dimensional scene according to claim 1, characterized in that: The determining whether there is a collision risk, performing automatic obstacle avoidance and dynamic route planning specifically includes: a. If the distance measurement value S(W i ) is greater than the preset first collision threshold, and the measured depth value h(W i ) is greater than the preset second collision threshold, it is determined that the UAV does not have a collision risk, and the UAV flies along the original flight path to the target point; b. If the distance measurement value S(W i ) is greater than the preset first collision threshold, and the measured depth value h(W i ) is less than or equal to the preset second collision threshold, the drone is judged to have a bottom collision risk. At this time, the drone stops moving forward and rises to the depth measured value h(W below the drone). i ) is greater than the preset second collision threshold, continue to fly to the previous target point; c. If the distance measurement value S(W i ) is less than or equal to the preset first collision threshold, and the measured depth value h(W i ) is greater than a preset second collision threshold, it is determined that the UAV has a front collision risk. At this time, the image in front of the UAV is collected and image processing is used to make obstacle avoidance decisions; d. If the distance measurement value S(W i ) is less than or equal to the preset first collision threshold, and the measured depth value h(W i ) is less than or equal to the preset second collision threshold, it is determined that the drone has a risk of collision with the bottom and the front. At this time, the drone stops moving forward and rises to the depth measured value h(W below the drone). i ) is greater than the preset second collision threshold, the image in front of the drone is collected and image processing is used to make obstacle avoidance decisions.

4. The method for dynamically planning a UAV inspection path in a three-dimensional scene according to claim 1, characterized in that: S1: obtaining an initial flight path according to an existing terrain map; specifically includes: S101: Perform three-dimensional environment modeling according to the existing topographic map to obtain a three-dimensional map; S102: Determine a plurality of flight path points W from the patrol area of ​​the three-dimensional map i , i represents the number of path points, and the initial flight path is obtained; S103: Mapping the initial flight path to the three-dimensional map and obtaining each flight path point W i The depth reference value H(W i ).

5. The method for dynamically planning a UAV inspection path in a three-dimensional scene according to claim 4, characterized in that: S102: Determine the flight path point W from the patrol area of ​​the three-dimensional map i , get the initial flight path; including: First, create a preliminary flight path from a 3D map based on the patrol area; Secondly, setting constraints and constructing an objective function, wherein the constraints include at least: a flight path cost constraint, an obstacle avoidance cost constraint and a flight altitude cost constraint, and the objective function includes: the shortest flight path and the smallest flight threat; Finally, multiple UAV path points W of the preliminary flight path are determined according to the constraint conditions and the objective function. i Position and get the initial flight path.

6. The method for dynamically planning a UAV inspection path in a three-dimensional scene according to claim 5, characterized in that: The constraints include at least: flight path cost constraint, obstacle avoidance cost constraint and flight altitude cost constraint, and the objective function includes: shortest flight path and minimum flight threat, specifically including: The flight path cost constraint: The flight path includes a plurality of flight path points W i , flight path point W i The coordinates are W i (x i ,y i , z i ), the distance between two adjacent flight path points is D(W i , W i+1 ); The total flight path cost is The obstacle avoidance cost constraint: Obtain the contour map of the inspection area. Assume that the measured value of the drone's flight altitude depth is h(W i ), determine the shortest distance s between the UAV and the contour line min , the corresponding contour lines are determined from the contour topographic map, and the width d outside the contour lines is set as the regional threat area; For any obstacle, the UAV is at a flight height of h(W i ), its risk cost R j It is expressed as: Among them, θ is the set first cost coefficient, j is the number of obstacle avoidance points during flight; The total risk cost is The flight altitude cost constraint: Set threat zone: above minimum altitude h min , below the minimum height h min + a height range of one unit height h0 (h min ,h min +h0), h0 is determined based on the height of local vegetation, and the minimum height h min Determined by terrain height in a three-dimensional map; Flight Waypoint W i The flight altitude cost h i for: Among them, ɡ is the set second cost coefficient; the maximum height h max The maximum flight altitude of the drone; The total height cost is The flight path is the shortest f1(A): The minimum flight threat is f2(A):

7. The method for dynamically planning a UAV inspection path in a three-dimensional scene according to claim 5, characterized in that: The multiple UAV path points W of the preliminary flight path are determined according to the constraint conditions and the objective function. i Position, get the initial flight path, including: Substitute the constraints and objective function into the MATLAB mathematical tool and use the fmincon function to solve and calculate the multiple UAV path points W of the preliminary flight path. i Position, a preliminary flight path can be obtained.

8. The method for dynamically planning a UAV inspection path in a three-dimensional scene according to claim 1, characterized in that: The reduced flight speed is reduced to 20%-50%.

9. A dynamic planning system for a drone inspection path in a three-dimensional scene, comprising a drone and a flight control device, characterized in that: The flight control device adopts a dynamic planning method for a drone inspection path in a three-dimensional scene as described in any one of claims 1 to 8.

10. A flight control device comprising: A processor and a memory storing instructions executable by the processor, wherein the processor executes the dynamic planning method of a drone inspection path in a three-dimensional scene as described in any one of claims 1 to 8.

Images