Method and system for autonomously planning flight path of road slope inspection unmanned aerial vehicle

Abstract

The invention relates to the technical field of flight path planning, in particular to an autonomous flight path planning method and system for a road slope inspection unmanned aerial vehicle, and the method comprises the following steps: controlling the unmanned aerial vehicle to fly in a road slope region according to a preset flight path, collecting the aerial image data of the unmanned aerial vehicle, analyzing the aerial inspection adaptive feature value of the unmanned aerial vehicle, and determining the flight path of the unmanned aerial vehicle. Judging whether the feature value is lower than a preset inspection adaptation feature threshold value or not; analyzing a field-of-view missing feature set of the unmanned aerial vehicle based on the aerial image data of the unmanned aerial vehicle if the field-of-view missing feature set is lower than the preset routing inspection adaptive feature value threshold; performing flight path planning processing on the unmanned aerial vehicle based on the field-of-view missing feature set; according to the method, the unmanned aerial vehicle is subjected to flight path planning processing on the basis of the field-of-view missing feature set, so that when an aerial picture is deviated, fuzzy or blind areas, judgment and automatic adjustment are performed in real time, self-adaptive optimization of the flight path is realized, and the road slope inspection efficiency is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of flight path planning technology, and more specifically, to a method and system for autonomous flight path planning of a highway slope inspection drone. Background Technology

[0002] Autonomous flight path planning technology for highway slope inspection drones is an important technology. Specifically applied in the scenario of safety monitoring and inspection of highway slopes, the core of which is to enable drones to autonomously plan flight paths in complex terrain environments and collect slope images in real time, thereby achieving visualized inspection of slope structural status, vegetation distribution, and the integrity of protective facilities.

[0003] Existing drone flight control technology for highway slope inspection faces several core problems in practical applications, including reliance on manual pre-setting for flight path planning, insufficient field of view coverage, and difficulty in timely re-shooting of blurred or missed areas. The key reason for these problems is that existing technologies mostly use fixed flight paths or regular grid scanning methods, without considering the dynamic matching relationship between image feedback during drone flight and the geometry of the slope. This makes it easy for drones to experience field of view shifts or blind spots in areas with large slope undulations, significant terrain obstruction, or strong wind disturbances. At the same time, existing systems usually only plan the path before the mission, lacking a dynamic judgment mechanism based on real-time aerial images. It is difficult to adjust the flight path according to features such as image clarity, edge integrity, and coverage completeness. As a result, once a shooting deviation occurs during drone inspection, the flight path can only be reset manually, which reduces inspection efficiency and increases the risk of mission interruption.

[0004] To address this technical challenge, we have developed a method and system for autonomous flight path planning for highway slope inspection drones. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for autonomous flight path planning of a highway slope inspection drone, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, one of the objectives of this invention is to provide an autonomous flight path planning method for a highway slope inspection drone, comprising the following steps:

[0007] S1. Control the drone to fly in the highway slope area according to the preset trajectory and collect aerial image data of the drone;

[0008] S2. Based on the aerial image data of the UAV, analyze the aerial inspection adaptation feature value of the UAV and determine whether it is lower than the preset inspection adaptation feature threshold.

[0009] S3. If the value is lower than the preset inspection adaptation feature value threshold, then analyze the missing feature set of the drone's field of view based on the drone's aerial image data.

[0010] S4. Perform trajectory planning for UAVs based on the feature set of missing field of view;

[0011] S5. If the value is not lower than the preset inspection adaptation characteristic value threshold, the drone will continue to fly in the highway slope area according to the preset flight path.

[0012] As a further improvement to this technical solution, the aerial image data specifically refers to the pixel value and two-dimensional coordinates of each pixel in the aerial image. The specific steps for analyzing the inspection adaptation of the UAV are as follows: Based on the aerial image data of the UAV, extract the inspection adaptation feature set of the UAV, including edge integrity feature value, shooting deviation feature value, and field of view coverage feature value; Based on the inspection adaptation feature set of the UAV, analyze the inspection adaptation feature value of the UAV.

[0013] As a further improvement to this technical solution, the specific steps for extracting the inspection adaptation feature set of the UAV are as follows: read the aerial image data of the UAV and preprocess it; perform feature analysis on the preprocessed aerial image data of the UAV to obtain the inspection adaptation feature set of the UAV.

[0014] As a further improvement to this technical solution, the specific formula for calculating the inspection adaptation characteristic value of the UAV is as follows: Where XjP is the inspection adaptation feature value of the UAV, ByW is the edge integrity feature value of the UAV, α1 is the edge integrity adjustment coefficient stored in the database, PsC is the shooting deviation feature value of the UAV, α2 is the shooting deviation adjustment coefficient stored in the database, FcG is the field of view coverage feature value of the UAV, α3 is the field of view coverage adjustment coefficient stored in the database, and α1+α2+α3=1.

[0015] As a further improvement to this technical solution, the specific steps for analyzing the missing feature set of the UAV's field of view are as follows: Spatial comparison processing is performed between the UAV's aerial image data and the slope target template image stored in the database to obtain the UAV's missing pixel set of the slope; Based on the UAV's missing pixel set of the slope, the missing feature set of the UAV's field of view is analyzed, including the missing area location feature value, the missing area feature value, and the missing direction feature value.

[0016] As a further improvement to this technical solution, the specific steps for obtaining the missing pixel set of the UAV slope are as follows: Spatial registration processing is performed on the aerial image data of the UAV and the slope target template image stored in the database; the spatially registered aerial image data of the UAV is compared with the slope target template image stored in the database to obtain the missing pixel set of the UAV; and the missing pixel set of the UAV is aggregated to obtain the missing pixel set of the UAV slope.

[0017] As a further improvement to this technical solution, the missing pixel set includes the pixel value and two-dimensional coordinates of each missing pixel. The specific steps of the aggregation process are as follows: Based on the pixel value and two-dimensional coordinates of each missing pixel of the UAV, analyze the Euclidean distance values ​​of several groups of adjacent pixels of the UAV; perform connectivity processing based on the Euclidean distance values ​​of each group of adjacent pixels of the UAV to obtain the missing pixel set of the UAV slope.

[0018] As a further improvement to this technical solution, the specific steps for analyzing the missing feature set of the UAV's field of view are as follows: Read the missing pixel set of the UAV's slope, which includes the pixels and two-dimensional coordinates of several missing pixels; Based on the pixels and two-dimensional coordinates of each missing pixel of the UAV, extract the missing region location feature value, missing area feature value, and missing direction feature value of the UAV.

[0019] As a further improvement to this technical solution, the specific steps for trajectory planning of the UAV based on the missing field-of-view feature set are as follows: Extract the UAV's trajectory offset vector based on the missing region position feature values; extract the UAV's trajectory adjustment magnitude based on the missing area feature values; extract the UAV's attitude adjustment angle based on the missing direction feature values; and perform trajectory correction processing on the UAV based on the trajectory offset vector, trajectory adjustment magnitude, and attitude adjustment angle.

[0020] The second objective of this invention is to provide an autonomous flight path planning system for a highway slope inspection drone, including any one of the above-described features, comprising: a data acquisition module for controlling the drone to fly within a highway slope area according to a preset flight path and collecting aerial image data of the drone; an aerial inspection adaptation judgment module for analyzing the drone's aerial inspection adaptation feature value based on the drone's aerial image data and determining whether it is lower than a preset inspection adaptation feature threshold; a field-of-view missing analysis module for analyzing the drone's field-of-view missing feature set based on the drone's aerial image data when the drone's aerial inspection adaptation feature value is lower than the preset inspection adaptation feature value threshold; a flight path planning feedback module for performing flight path planning processing on the drone based on the field-of-view missing feature set; and a flight path holding control module for continuing to control the drone to fly within the highway slope area according to the preset flight path when the drone's aerial inspection adaptation feature value is not lower than the preset inspection adaptation feature value threshold.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0022] This invention acquires aerial image data of a drone flying over a highway slope area in real time, generates aerial inspection adaptation feature values ​​based on the aerial image data, and determines whether the values ​​are below a preset inspection adaptation feature threshold. When the threshold is below, it automatically calls up the slope target template image corresponding to the drone's current position, extracts the missing feature set of the field of view, and generates a trajectory correction scheme based on it. Otherwise, it maintains the drone's predetermined trajectory and attitude to ensure continuous execution of the flight mission. Thus, when there are offsets, blurs, or blind spots in the aerial images, it can immediately judge and automatically adjust, effectively avoiding mission interruptions and repeated flights caused by manual replanning of the flight path. It achieves adaptive trajectory optimization and complete coverage of the inspection field of view, thereby significantly improving the efficiency of highway slope inspection. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating the overall workflow of the present invention;

[0024] Figure 2 This is a schematic diagram of the overall structure of the present invention;

[0025] The meanings of the labels in the diagram are as follows:

[0026] 1. Data acquisition module; 2. Aerial inspection and adaptation judgment module; 3. Field of view gap analysis module; 4. Track planning feedback module; 5. Track holding control module. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0028] Example 1

[0029] Please see Figure 1 As shown, one of the objectives of this embodiment is to provide an autonomous flight path planning method for a highway slope inspection UAV, including the following steps:

[0030] S1. Control the drone to fly within the highway slope area according to the preset flight path, and collect aerial image data of the drone in real time during the flight.

[0031] S2. Based on the aerial image data of the UAV, analyze the aerial inspection adaptation feature value of the UAV and determine whether it is lower than the preset inspection adaptation feature threshold. The aerial image data specifically consists of the pixel value and two-dimensional coordinates of each pixel in the aerial image. The specific steps for analyzing the inspection adaptation of the UAV are as follows: Based on the aerial image data of the UAV, extract the inspection adaptation feature set of the UAV, including edge integrity feature value, shooting deviation feature value, and field of view coverage feature value; Based on the inspection adaptation feature set of the UAV, analyze the inspection adaptation feature value of the UAV.

[0032] The specific steps for extracting the inspection adaptation feature set of the UAV are as follows: Read the aerial image data of the UAV and perform preprocessing, such as noise suppression (median filtering or bilateral filtering can be used to remove random noise), brightness equalization (historical equalization can be used to adjust the overall brightness distribution of the image), and edge enhancement (gradient operations or sharpening filters can be used to highlight edge structures); perform feature analysis on the preprocessed UAV aerial image data to obtain the inspection adaptation feature set of the UAV, which is as follows:

[0033] Image region segmentation is performed based on image grayscale distribution: The grayscale gradient change rate is used as the segmentation criterion to extract continuous regions with significant grayscale gradient changes. In aerial images, regions with significant grayscale gradient changes typically correspond to areas with obvious terrain undulations or strong color contrasts. The grayscale gradient change is most concentrated along the longitudinal direction in highway slope areas, thus allowing for the identification of the slope region's pixel set. Within the identified slope region, edge detection operators (such as the Sobel or Canny operators) are used to extract edge pixel sets pixel by pixel. Contour tracking is performed on the detected edge pixels, connecting continuous edges point by point based on pixel spatial adjacency relationships to form closed contours. The ratio of the number of continuous edge pixels in the closed contour to the theoretical number of edge pixels in the region is calculated and used as the edge integrity feature value. A higher feature value indicates a more continuous and clearer slope boundary, which can be used to characterize the integrity of the slope boundary in the current aerial image.

[0034] The principal direction distribution of pixel grayscale in the slope area of ​​the image is calculated using grayscale projection or structural orientation filtering, i.e., the rate of grayscale change along the horizontal and vertical directions. The direction with the largest rate of grayscale change is selected as the principal direction line of the slope surface. Simultaneously, the camera pitch and roll angle information recorded by the UAV attitude sensor (inertial measurement unit IMU) at the moment of shooting is read to determine the UAV camera's line of sight direction. The angle between the principal direction line and the camera's line of sight direction is calculated to obtain the shooting deviation angle at that moment. The absolute value of this angle is used as the shooting deviation feature value to characterize the degree of deviation between the UAV's shooting viewpoint and the slope normal direction. The smaller the shooting deviation feature value, the more perpendicular the UAV's viewpoint is to the slope surface, the smaller the image geometric distortion, and the higher the detail fidelity.

[0035] Based on grayscale thresholding or color segmentation methods, the image is divided into three pixel sets: sky region, road region, and slope region. The slope region is usually represented as a continuous region with a high grayscale gradient. For the identified slope region pixel set, the number of pixels is counted and recorded as , and the total number of pixels in the entire image is also counted. The ratio of the two is calculated and used as the field of view coverage feature value. This feature value is used to characterize the coverage ratio of the target slope in the current image captured by the UAV. When the field of view coverage feature value is high, it indicates that the target area occupies a large proportion of the image and is fully covered. When the value is low, it indicates that there are blind spots or incomplete coverage. The edge integrity feature value, shooting deviation feature value, and field of view coverage feature value are activated by the Sigmoid function, and their corresponding results are mapped between 0 and 1.

[0036] The specific formula for calculating the inspection adaptation characteristic value of the UAV is as follows: Where XjP is the inspection adaptation feature value of the UAV, ByW is the edge integrity feature value of the UAV, α1 is the edge integrity adjustment coefficient stored in the database, PsC is the shooting deviation feature value of the UAV, α2 is the shooting deviation adjustment coefficient stored in the database, FcG is the field of view coverage feature value of the UAV, α3 is the field of view coverage adjustment coefficient stored in the database, and α1+α2+α3=1.

[0037] It should be noted that the steps for obtaining the edge integrity adjustment coefficient α1, shooting deviation adjustment coefficient α2, and field of view coverage adjustment coefficient α3 stored in the database are as follows: Obtain the edge integrity feature values, shooting deviation feature values, and field of view coverage feature values ​​from several historical drone tests. Extract the mean values ​​of these three features respectively, and sum them to obtain the adaptation sum. Ratio the mean values ​​of these features to the adaptation sum, and use the results as the edge integrity adjustment coefficient α1, shooting deviation adjustment coefficient α2, and field of view coverage adjustment coefficient α3.

[0038] S3. If the value is lower than the preset inspection adaptation feature value threshold, then analyze the missing feature set of the drone's field of view based on the drone's aerial image data.

[0039] The specific steps for analyzing the missing feature set of the UAV's field of view are as follows: The UAV's aerial image data is spatially compared with the slope target template images (located at the UAV's current position) stored in the database. (These template images are images from a slope area template library established during system initialization based on historical inspection data and GIS data. Each template image corresponds one-to-one with the geographic coordinates, slope number, and elevation reference data of its shooting area. When the UAV acquires aerial images during inspection, the system automatically retrieves the most matching slope target template image from the database based on the UAV's real-time positioning information, including latitude, longitude, altitude, and yaw angle parameters.) This yields the UAV's missing pixel set for the slope. Based on the UAV's missing pixel set for the slope, the missing feature set of the UAV's field of view is analyzed, including the missing area location feature value, missing area feature value, and missing direction feature value.

[0040] The specific steps to obtain the missing pixel set of the slope from the UAV are as follows:

[0041] Spatial registration is performed between the drone's aerial imagery and the slope target template images stored in the database. Specifically: First, the geographic identification information and attitude parameters of the currently captured drone image are read. The geographic identification information includes the latitude and longitude coordinates and altitude data at the time of capture, while the attitude parameters include the drone's pitch, yaw, and roll angles. Simultaneously, the slope target template image corresponding to this latitude and longitude range is retrieved from the database, and its spatial reference information is read. Second, feature point extraction is performed on both images: Methods based on Scale Invariant Feature Transform (SIFT) or Speed-Up Robust Feature Transform (SURF) are used to extract ground feature points from both the aerial image and the template image. The aerial image is constructed by first identifying two sets of features, including the two-dimensional coordinate features of slope contours, road edges, and obvious terrain texture points. Then, two sets of feature points are matched, and the correspondence between each feature point in the aerial image and the template image is calculated using the Euclidean distance minimum matching principle. Next, based on the matched feature point pairs, the spatial affine transformation matrix or perspective transformation matrix is ​​solved using the least squares method. The transformation matrix describes the geometric relationship between the aerial image and the template image in the dimensions of translation, rotation, and scaling. Finally, the obtained transformation matrix is ​​used to perform geometric correction on the aerial image, mapping the two-dimensional coordinates of each pixel in the aerial image to the coordinate system of the template image, thus achieving precise alignment between the aerial image and the template image in the same spatial reference system.

[0042] The aerial image data of the UAV after spatial registration is compared with the slope target template image stored in the database to obtain the missing pixel set of the UAV. Specifically, the spatially registered aerial image and the slope target template image are compared pixel by pixel according to a unified spatial coordinate system. For pixels at the same coordinate position in the two images, their gray value or color component value (including red, green and blue three-channel components) is read, and the gray value difference or color difference of the corresponding pixels is calculated. Then, a pixel difference threshold is set. When the gray value difference or color difference at a certain coordinate position is greater than the threshold, or when there is a valid pixel at the coordinate position in the template image but an invalid pixel (e.g., the brightness value is zero or a blank pixel) at the corresponding position in the aerial image, the pixel is determined to be a missing pixel. All missing pixels are counted to form the missing pixel set of the UAV. The missing pixel set is used to represent the slope area that is not covered or is abnormally captured in the current aerial image.

[0043] The missing pixel set of the UAV is then aggregated to obtain the missing pixel set of the UAV slope.

[0044] The missing pixel set includes the pixel value and two-dimensional coordinates of each missing pixel. The specific steps of the aggregation process are as follows: Based on the pixel value and two-dimensional coordinates of each missing pixel of the UAV, analyze the Euclidean distance values ​​of several groups of adjacent pixels of the UAV; perform connectivity processing based on the Euclidean distance values ​​of each group of adjacent pixels of the UAV to obtain the missing pixel set of the UAV slope, specifically: spatially sort each missing pixel in the missing pixel set according to its two-dimensional coordinate values, and calculate the Euclidean distance between any adjacent pixels in a unified coordinate system; set a pixel adjacency threshold, and when the Euclidean distance between adjacent pixels is less than the threshold, determine that the two pixels belong to the same family. In the same missing region, starting from any missing pixel, a layer-by-layer search is performed along the direction of its adjacent pixels, grouping all missing pixels that meet the adjacency condition into the same connected region until no new adjacent points are added, and defining this connected region as an initial missing region; next, all initial missing regions are filtered by area, and the number of pixels contained in each missing region is calculated. When the number of pixels is less than a preset minimum area threshold, the region is determined to be an isolated noise region and is removed; finally, the pixel sets of all the missing regions that have passed the filtering are merged to form the slope missing pixel set of the UAV, which includes several missing pixels.

[0045] The specific steps for analyzing the missing feature set of the UAV's field of view are as follows: Read the missing pixel set of the UAV's slope, which includes the pixels and two-dimensional coordinates of several missing pixels; based on the pixels and two-dimensional coordinates of each missing pixel of the UAV, extract the location feature value, area feature value, and direction feature value of the missing region of the UAV, specifically as follows:

[0046] The missing pixels in the slope missing pixel set are clustered according to spatial adjacency, and those with an adjacent distance less than a preset adjacency threshold D are grouped together. t The missing pixels are divided into the same missing region; for all missing pixels in each missing region, their x and y coordinates are extracted, the average of the x and y coordinates is calculated, and the two are combined to form the center coordinate point of the missing region; the Euclidean distance between the center coordinate point and the center point of the entire aerial image is calculated, and the distance value is defined as the position feature value of the missing region, which is used to characterize the spatial offset of the missing region relative to the center of the drone's field of view.

[0047] The number of missing pixels is counted, and the total number of pixels in the entire aerial image is counted. The ratio is then used as the feature value of the missing area. When the ratio is large, it indicates that there is a large area of ​​uncovered area in the aerial image. When the ratio is small, it indicates that the missing area is small and the coverage is relatively sufficient.

[0048] Using the center coordinates of each missing region and the center point of the aerial image as endpoints, calculate the direction vector of the line connecting the two points; using the horizontal axis of the aerial image as a reference direction, calculate the angle between the direction vector and the horizontal axis; define this angle as the missing direction feature value, which is used to characterize the orientation of the missing region within the image's field of view; when the missing direction feature value points upwards in the aerial image, it indicates that the drone's shooting angle is too low; when the missing direction feature value points to the left or right side of the image, it indicates that the drone's field of view has a lateral shift.

[0049] S4. Perform trajectory planning for UAVs based on the feature set of missing field of view;

[0050] The specific steps for UAV trajectory planning based on the missing field-of-view feature set are as follows:

[0051] Based on the location feature values ​​of the missing regions of the UAV, the UAV's trajectory offset vector is extracted. Specifically, spatial direction analysis is performed on each location feature value of the missing region, and the direction vector component corresponding to the feature value is extracted. Taking the center point of the aerial image as the origin, the direction of the line connecting the center coordinates of the missing region and the center coordinates of the image is defined as the offset direction. The unit vector of the connecting direction is multiplied with the location feature value of the missing region (i.e., Euclidean distance) to obtain the UAV's trajectory offset vector, which is used to characterize the displacement direction and distance that the UAV should move along in this direction.

[0052] Based on the missing area feature value of the UAV, the UAV's trajectory adjustment range is extracted. Specifically, the missing area feature value of each missing region is read, and its ratio to a preset coverage area threshold is calculated. When the missing area feature value is higher than the threshold, it indicates that there is a large uncovered area in the aerial image. At this time, the magnitude of the trajectory offset vector is increased according to the ratio, increasing the flight offset distance of the UAV. When the missing area feature value is lower than the threshold, it indicates that the uncovered area is small. At this time, the magnitude of the trajectory offset vector is reduced proportionally to reduce the correction range and maintain the continuity of the image. The corrected magnitude is defined as the UAV's trajectory adjustment range, which is used to control the scale of the UAV's movement distance during the trajectory correction process.

[0053] Based on the missing directional feature value of the UAV, the attitude adjustment angle of the UAV is extracted. Specifically, the azimuth angle corresponding to the missing directional feature value is used as a reference to calculate the angle difference between the azimuth angle and the current camera's line of sight of the UAV. When the angle difference is positive, the UAV gimbal is controlled to rotate clockwise around the vertical axis by the absolute value of the angle difference. When the angle difference is negative, the gimbal is controlled to rotate counterclockwise around the vertical axis by the absolute value of the angle difference. If the missing directional feature value points to the top or bottom of the image, the pitch angle of the gimbal is further adjusted to make the camera's line of sight aligned with the center direction of the missing area, ensuring that the reshot image is continuous with the field of view boundary of the original image.

[0054] The UAV undergoes trajectory correction processing based on trajectory offset vector, trajectory adjustment range, and attitude adjustment angle. Specifically, the directional component of the trajectory offset vector, the distance component of the trajectory adjustment range, and the rotation component of the attitude adjustment angle are combined to generate a trajectory correction scheme for the UAV. The trajectory correction scheme is then sent to the UAV's flight control system, which controls the UAV to re-execute the aerial photography mission according to the corrected flight parameters. During execution, the UAV's flight position and camera attitude status are monitored in real time. When the aerial footage reaches the preset coverage standard, the trajectory correction action is automatically terminated and the UAV returns to the original inspection path, thereby achieving accurate re-shooting of missing areas and adaptive trajectory adjustment.

[0055] S5. If the value is not lower than the preset inspection adaptation characteristic value threshold, the drone will continue to fly in the highway slope area according to the preset flight path.

[0056] The second objective of this invention is to provide an autonomous flight path planning system for a highway slope inspection drone that achieves any of the above-mentioned objectives. The system includes: a data acquisition module for controlling the drone to fly within a highway slope area according to a preset flight path and collecting aerial image data of the drone in real time during flight; an aerial inspection adaptation judgment module for analyzing the drone's aerial inspection adaptation feature value based on the drone's aerial image data and determining whether it is lower than a preset inspection adaptation feature threshold; a field-of-view missing analysis module for analyzing the drone's field-of-view missing feature set based on the drone's aerial image data when the drone's aerial inspection adaptation feature value is lower than the preset inspection adaptation feature value threshold; a flight path planning feedback module for performing flight path planning processing on the drone based on the field-of-view missing feature set; and a flight path holding control module for continuing to control the drone to fly within the highway slope area according to the preset flight path when the drone's aerial inspection adaptation feature value is not lower than the preset inspection adaptation feature value threshold.

[0057] This invention controls a drone to fly within a highway slope area according to a preset flight path, collects aerial image data from the drone, analyzes the drone's aerial inspection adaptation feature value, and determines whether it is lower than a preset inspection adaptation feature threshold. If it is lower than the preset inspection adaptation feature value threshold, the invention analyzes the drone's field-of-view missing feature set based on the drone's aerial image data and performs flight path planning processing on the drone based on the field-of-view missing feature set. Otherwise, the invention continues to control the drone to fly within the highway slope area according to the preset flight path.

[0058] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for autonomous trajectory planning of a highway slope inspection UAV, characterized in that: Includes the following steps: S1. Control the drone to fly within the highway slope area according to the preset flight path and collect aerial image data of the drone; S2. Based on the aerial image data of the UAV, analyze the aerial inspection adaptation feature value of the UAV and determine whether it is lower than the preset inspection adaptation feature threshold. S3. If the value is lower than the preset inspection adaptation feature value threshold, then analyze the missing feature set of the drone's field of view based on the drone's aerial image data. S4. Perform trajectory planning for UAVs based on the feature set of missing field of view; S5. If the value is not lower than the preset inspection adaptation characteristic value threshold, the drone will continue to fly in the highway slope area according to the preset flight path.

2. The method for autonomous trajectory planning of a highway slope inspection UAV according to claim 1, characterized in that: The aerial image data specifically refers to the pixel value and two-dimensional coordinates of each pixel in the aerial image. The specific steps for analyzing the inspection adaptation of the UAV are as follows: Based on the aerial image data of UAVs, the inspection adaptation feature set of UAVs is extracted, including edge integrity feature value, shooting deviation feature value, and field of view coverage feature value. Based on the inspection adaptation feature set of UAVs, we analyze the inspection adaptation feature values ​​of UAVs.

3. The method for autonomous trajectory planning of a highway slope inspection UAV according to claim 2, characterized in that: The specific steps for extracting the inspection adaptation feature set of UAVs are as follows: Read aerial image data from the drone and perform preprocessing; Feature analysis was performed on the preprocessed aerial image data of the UAV to obtain the inspection adaptation feature set of the UAV.

4. The method for autonomous trajectory planning of a highway slope inspection UAV according to claim 2, characterized in that: The specific formula for calculating the inspection adaptation characteristic value of the UAV is as follows: Where XjP is the inspection adaptation feature value of the UAV, ByW is the edge integrity feature value of the UAV, α1 is the edge integrity adjustment coefficient stored in the database, PsC is the shooting deviation feature value of the UAV, α2 is the shooting deviation adjustment coefficient stored in the database, FcG is the field of view coverage feature value of the UAV, α3 is the field of view coverage adjustment coefficient stored in the database, and α1+α2+α3=1.

5. The method for autonomous trajectory planning of a highway slope inspection UAV according to claim 1, characterized in that: The specific steps for analyzing the missing feature set of a UAV's field of view are as follows: The drone's aerial image data is spatially compared with the slope target template images stored in the database to obtain the drone's missing pixel set for the slope. Based on the missing pixel set of the slope from the UAV, we analyze the missing feature set of the UAV's field of view, including the location feature value of the missing region, the missing area feature value, and the missing direction feature value.

6. The method for autonomous trajectory planning of a highway slope inspection UAV according to claim 5, characterized in that: The specific steps to obtain the missing pixel set of the slope from the UAV are as follows: Spatial registration processing is performed between the drone's aerial image data and the slope target template images stored in the database; The aerial image data of the UAV that has undergone spatial registration processing is compared with the slope target template image stored in the database to obtain the set of missing pixels of the UAV. The missing pixel set of the UAV is then aggregated to obtain the missing pixel set of the UAV slope.

7. The method for autonomous trajectory planning of a highway slope inspection UAV according to claim 5, characterized in that: The set of missing pixels includes the pixel value and two-dimensional coordinates of each missing pixel. The specific steps of the aggregation process are as follows: Based on the pixel value and two-dimensional coordinates of each missing pixel of the drone, the Euclidean distance values ​​of several groups of adjacent pixels of the drone are analyzed. The missing pixel set of the slope is obtained by performing connectivity processing based on the Euclidean distance value of each group of adjacent pixels of the UAV.

8. The method for autonomous trajectory planning of a highway slope inspection UAV according to claim 5, characterized in that: The specific steps for analyzing the missing feature set of a UAV's field of view are as follows: Read the set of missing pixels on the slope from the drone, which includes the pixels and two-dimensional coordinates of several missing pixels; Based on the pixel and two-dimensional coordinates of each missing pixel in the UAV, the location feature value, area feature value, and direction feature value of the missing region of the UAV are extracted.

9. The method for autonomous trajectory planning of a highway slope inspection UAV according to claim 5, characterized in that: The specific steps for UAV trajectory planning based on the missing field-of-view feature set are as follows: Based on the location feature values ​​of the missing region of the UAV, the UAV's trajectory offset vector is extracted; Based on the missing area feature value of the UAV, the trajectory adjustment range of the UAV is extracted; Based on the missing directional feature values ​​of the UAV, the attitude adjustment angle of the UAV is extracted; The UAV's trajectory is corrected based on the trajectory offset vector, trajectory adjustment range, and attitude adjustment angle.

10. A trajectory autonomous planning system for implementing a highway slope inspection UAV according to any one of claims 1-9, characterized in that, include: The data acquisition module is used to control the drone to fly in the highway slope area according to the preset trajectory and collect aerial image data of the drone; The aerial inspection adaptation judgment module is used to analyze the aerial inspection adaptation feature value of the UAV based on the UAV's aerial image data, and determine whether it is lower than the preset inspection adaptation feature threshold. The field-of-view missing analysis module is used to analyze the field-of-view missing feature set of the UAV based on the UAV's aerial image data when the aerial inspection adaptation feature value of the UAV is lower than the preset inspection adaptation feature value threshold. The trajectory planning feedback module is used to perform trajectory planning processing on UAVs based on the missing feature set of the field of view; The trajectory-keeping control module is used to continue controlling the drone to fly within the highway slope area according to the preset trajectory if the drone's aerial inspection adaptation feature value is not lower than the preset inspection adaptation feature value threshold.

Images