Path planning and defect marking method, system and terminal for supervising inspection of unmanned aerial vehicle

Abstract

The invention relates to a path planning and defect marking method, system and terminal for patrol inspection of a supervision unmanned aerial vehicle, and belongs to the technical field of railway engineering supervision, and the path planning and defect marking method comprises the steps: obtaining a grid safety map of a target patrol inspection area; based on the grid security map, generating an original inspection path for supervising the inspection of the unmanned aerial vehicle; when the supervision unmanned aerial vehicle performs patrol inspection along the original patrol inspection path, patrol inspection process information is obtained; performing defect marking on the grid safety map according to the inspection process information and the flight position of the supervision unmanned aerial vehicle; and dynamically adjusting the inspection path according to the defect grids marked by the defects. The method has the beneficial effect of improving the overall inspection efficiency of the unmanned aerial vehicle.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of railway engineering supervision, in particular to a path planning and defect labeling method, system and terminal for supervision unmanned aerial vehicle inspection. BACKGROUND

[0002] With the rapid development of railway construction in China, the coverage of railway lines is continuously expanding, and the construction and operation management of railway engineering are facing greater and greater challenges. The safety and quality of railway engineering are directly related to the safety and efficiency of railway transportation, so it is necessary to comprehensively, timely and accurately supervise the railway engineering. Traditional railway engineering supervision mainly relies on manual inspection and ground monitoring equipment, however, these methods have low efficiency, limited coverage, and difficulty in finding hidden defects. With the rapid development of unmanned aerial vehicle technology, its application in railway engineering supervision is attracting more and more attention. Unmanned aerial vehicles can quickly and efficiently obtain relevant information of railway engineering, providing new means and methods for railway engineering supervision.

[0003] Currently, some railway engineering supervision has begun to try to use unmanned aerial vehicles for inspection, using unmanned aerial vehicles equipped with high-definition cameras and optical sensors to fly along the railway track to obtain surface condition information of the track.

[0004] However, in practice, it is found that the unmanned aerial vehicle inspection path lacks reasonable planning, and most of them fly according to the preset fixed route, which leads to insufficient inspection of some key areas and repeated inspection of some areas that have been confirmed to be safe, resulting in low overall inspection efficiency. SUMMARY

[0005] In order to improve the overall inspection efficiency of unmanned aerial vehicles, the present application provides a path planning and defect labeling method, system and terminal for supervision unmanned aerial vehicle inspection.

[0006] In a first aspect, the present application provides a path planning and defect labeling method for supervision unmanned aerial vehicle inspection, which adopts the following technical scheme: A path planning and defect labeling method for supervision unmanned aerial vehicle inspection, comprising: obtaining a grid safety map of a target inspection area; generating an original inspection path for supervision unmanned aerial vehicle inspection based on the grid safety map; obtaining inspection process information when the supervision unmanned aerial vehicle inspects along the original inspection path; performing defect labeling on the grid safety map according to the inspection process information and the flight position of the supervision unmanned aerial vehicle; and dynamically adjusting the inspection path according to the defect grid of the defect labeling.

[0007] By adopting the technical scheme, the grid safety map of the target inspection area is acquired and the original inspection path is generated, so that the supervisor unmanned aerial vehicle can perform inspection according to the preset reasonable path, blind flight is avoided, unnecessary flight time and distance are reduced, compared with the traditional unplanned inspection mode, the efficiency of the inspection is greatly improved. The inspection path is dynamically adjusted according to the defect grid marked with defects, so that the unmanned aerial vehicle can perform key inspection on the area where defects are found, instead of continuing to inspect the area which has been confirmed to have no problem according to the original path, so that the inspection resources can be more targetedly allocated, and the overall inspection efficiency is improved.

[0008] Optionally, the step of acquiring the grid safety map of the target inspection area comprises: receiving topographic data, meteorological data and construction progress information of the target inspection area; generating a blank grid safety map of the target inspection area according to the topographic data; extracting slope data in the topographic data; determining a flight radius according to the meteorological data; mapping the construction progress information into a grid weight; according to the slope data, the flight radius and the grid weight, giving each grid in the blank grid safety map a corresponding flight state to generate a final grid safety map.

[0009] By adopting the technical scheme, the topographic data, meteorological data and construction progress information of the target inspection area are received, and these information is used to generate the grid safety map, which can comprehensively consider various factors affecting the flight safety of the unmanned aerial vehicle. According to various data, each grid in the blank grid safety map is given a corresponding flight state, so that the unmanned aerial vehicle can understand the flight safety of each area before performing the inspection task. The unmanned aerial vehicle can plan the flight path in advance according to the grid safety map, avoid the grid area marked as dangerous, reduce the flight time and possibility in the dangerous area, and further improve the safety in the inspection process.

[0010] Optionally, the step of generating the original inspection path of the supervisor unmanned aerial vehicle based on the grid safety map comprises: receiving node information of an inspection starting point, an inspection ending point and an inspection process node of the supervisor unmanned aerial vehicle; generating all flyable paths in the grid safety map according to the node information; calculating the passing cost corresponding to all flyable paths; selecting the flyable path corresponding to the minimum passing cost as the original inspection path.

[0011] By adopting the technical scheme, the inspection starting point, the inspection ending point and the inspection process node information of the supervisory unmanned aerial vehicle are received, so that the actual inspection requirements and targets are fully considered in path planning. The key information reflects the areas and the inspection sequence that need to be focused on in railway engineering supervision, so that the generated path can be more in line with the actual inspection task, blind path planning is avoided, and the fitting degree of the path planning and the actual requirements is improved. Through the quantitative calculation of the passing cost of the flyable path, the advantages and disadvantages of different paths can be objectively compared; the path with the minimum passing cost can effectively reduce the economic cost and resource consumption in the inspection process, and further improve the inspection efficiency.

[0012] Optionally, the step of calculating the passing cost corresponding to each flyable path comprises: obtaining the Euclidean distance between adjacent flight nodes of each flyable path; obtaining the flight weight of the corresponding grid of each flyable path; obtaining the historical selection times of the inspection starting point, the inspection ending point and the inspection process node; generating a node weight according to the historical selection times; calculating the passing cost corresponding to each flyable path according to the Euclidean distance, the flight weight and the node weight.

[0013] By adopting the technical scheme, the Euclidean distance between adjacent flight nodes, the flight weight of the corresponding grid, the historical selection times of the inspection starting point, the ending point and the process node are comprehensively considered when calculating the passing cost. The Euclidean distance reflects the physical length of the path, the flight weight reflects the environmental complexity, the obstacle situation and the like on the path, and the node weight considers the historical selection situation. The comprehensive consideration of these factors can more truly reflect the actual passing difficulty and cost of each path. The multi-dimensional consideration makes the evaluation of the path passing cost more comprehensive and accurate. In addition, different inspection tasks may have different emphases on various factors. The method can better meet the actual requirements of the railway engineering supervision unmanned aerial vehicle inspection by comprehensively considering various factors.

[0014] Optionally, the step of dynamically adjusting the inspection path according to the defect-labeled defect grid comprises: generating a severity weight of the defect grid according to the defect labeling; performing inspection sorting on the defect grid according to the severity weight; adjusting the original inspection path according to the inspection sorting, the inspection starting point and the inspection ending point information, and generating a specified inspection path.

[0015] By adopting the technical scheme, the severity of different defects can be quantified by generating the severity weight of the corresponding defect grid according to the defect labeling, so that the defect grid corresponding to the defect with high severity can be given priority in subsequent inspection sorting and path adjustment, so that a more reasonable inspection sequence can be planned for the unmanned aerial vehicle, and the unmanned aerial vehicle can first deal with the most urgent and critical problems. According to the inspection sorting, the inspection starting point and the inspection ending point information, the original inspection path is adjusted to generate a specified inspection path, so that the unmanned aerial vehicle can be prevented from flying in unnecessary areas, the flight time and distance are reduced, and the newly discovered defects can be inspected in a targeted manner in a timely manner.

[0016] Optionally, the step of adjusting the original inspection path according to the inspection sorting, the inspection starting point and the inspection ending point information to generate a specified inspection path comprises: The inspection sorting comprises high risk, medium risk and low risk; According to the inspection starting point and the high risk sorting, a first specified inspection path is generated; When the inspection reaches the end point of the first specified inspection path, a first inspection starting point of the medium risk sorting is determined; According to the first inspection starting point and the medium risk sorting, a second specified inspection path is generated; When the inspection reaches the end point of the second specified inspection path, a second inspection starting point of the low risk sorting is determined; According to the second inspection starting point, the low risk sorting and the inspection starting point, a third specified inspection path is generated; The inspection frequency of the first specified inspection path is greater than the inspection frequency of the second specified inspection path, which is greater than the inspection frequency of the third specified inspection path.

[0017] By adopting the technical scheme, the inspection path is planned in the order of high risk, medium risk and low risk, so that the high-risk areas and equipment can be processed first. The limited inspection resources are concentrated in the places that need most attention, and excessive time and effort are avoided in low-risk areas, so that the entire inspection process is more efficient. The inspection paths of different risk levels are set with different inspection frequencies, the high-risk path has the most inspection frequencies, and the low-risk path has the least inspection frequencies, which conforms to the principle of reasonable allocation of resources. The input of inspection resources can be flexibly adjusted according to the risk degree, so that manpower, material resources and financial resources can be optimally configured.

[0018] Optionally, the inspection path planning and defect labeling method further comprises: When the supervision unmanned aerial vehicle inspects along the specified inspection path, the percentage of the power of the supervision unmanned aerial vehicle is obtained; When the percentage of the power is lower than the set inspection power value, a return inspection path is generated according to the current power percentage, the inspection progress and the set minimum inspection power.

[0019] By adopting the technical scheme, the return inspection path is generated according to the current percentage of electric quantity, the inspection progress and the set minimum inspection electric quantity, which helps to reasonably plan the inspection task. When the electric quantity is insufficient, the unmanned aerial vehicle can timely adjust the path to return, thereby avoiding interruption or repeated inspection due to the electric quantity problem.

[0020] In a second aspect, the application provides a path planning and defect labeling system for supervising unmanned aerial vehicle inspection, which adopts the following technical scheme: A path planning and defect labeling system for supervising unmanned aerial vehicle inspection, comprising: a map acquisition module configured to acquire a grid safety map of a target inspection area; a path generation module configured to generate an original inspection path for supervising unmanned aerial vehicle inspection based on the grid safety map; an information acquisition module configured to acquire inspection process information when the supervising unmanned aerial vehicle inspects along the original inspection path; and a defect labeling module configured to label defects on the grid safety map according to the inspection process information and the flight position of the supervising unmanned aerial vehicle; a path adjustment module configured to dynamically adjust the inspection path according to the defect grid labeled by the defect labeling module.

[0021] In a third aspect, the application provides a terminal, which adopts the following technical scheme: A terminal, comprising: a memory storing a path planning and defect labeling program for supervising unmanned aerial vehicle inspection; a processor configured to execute the program stored on the memory to implement the steps of the path planning and defect labeling method for supervising unmanned aerial vehicle inspection.

[0022] In summary, the application has at least the following beneficial effects: By acquiring the grid safety map of the target inspection area and generating the original inspection path, the supervising unmanned aerial vehicle can inspect according to the pre-set reasonable path, thereby avoiding blind flight and reducing unnecessary flight time and distance. Compared with the traditional unplanned inspection method, the efficiency of the inspection is greatly improved. The inspection path is dynamically adjusted according to the defect grid labeled by the defect labeling module, so that the unmanned aerial vehicle can focus on the inspection of the area where defects are found, instead of continuing to inspect the areas that have been confirmed to have no problems according to the original path. In this way, the inspection resources can be more targetedly allocated, and the overall inspection efficiency is improved. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 is a first flowchart of an embodiment of the application; Figure 2 is a second flowchart of an embodiment of the application; Figure 3 This is a third flow chart of an embodiment of the present application; Figure 4 This is a fourth flow chart of an embodiment of the present application; Figure 5 This is a fifth flow chart of an embodiment of the present application; Figure 6 This is the sixth flow chart of the embodiment of the present application. DETAILED DESCRIPTION

[0024] To make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the following will be combined with the appended drawings of the embodiments of the present invention. Figure 1 -Attached Figure 6 The technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0025] The first embodiment of the present application discloses a method for path planning and defect marking for a supervisory drone inspection. Figure 1 , the path planning and defect marking method includes S110-S150: S110, obtaining a grid safety map of the target inspection area; S120, based on the grid safety map, generates the original inspection path for the supervisory drone inspection; S130, when the supervisory drone inspects along the original inspection route, obtaining inspection process information; S140, based on the inspection process information and the flight position of the supervision drone, defects are marked on the grid safety map; S150, dynamically adjusting the inspection path according to the defect grid marked with the defect.

[0026] Reference Figure 2 For S110, the specific steps of obtaining the grid safety map of the target inspection area may include S210-S260: S210, receiving terrain data, meteorological data and construction progress information of the target inspection area; S220, generating a blank grid safety map of the target inspection area based on the terrain data; S230, extracting slope data from the terrain data; S240, determining the flight radius based on meteorological data; S250, mapping the construction progress information into grid weights; S260, according to the slope data, flight radius and grid weight, the flight state corresponding to each grid in the blank grid safety map is given to generate the final grid safety map.

[0027] Specifically, the DEM (Digital Elevation Model) data of the target inspection area can be loaded through the GIS platform (such as ArcGIS), including latitude and longitude coordinates and corresponding altitude information. The real-time meteorological parameters such as wind speed can be obtained by calling the meteorological API (such as OpenWeatherMap) or can also be detected by the meteorological sensor installed in the target inspection area. The structured data (such as Excel table) of the construction area, progress percentage and other structured data of the target inspection area can be exported through the project management system. The latitude and longitude coordinates are converted into a regular grid using the meshgrid function of MATLAB, and a blank grid matrix is created, and the grid unit size can be set to 10m x 10m (adjustable according to the accuracy requirement). The slope data can be extracted with the help of ArcGIS tools, for example, in the 【3D Analyst tool】, execute the

slope

[0028] According to the mapping rule of the horizontal component of wind speed and the flight radius, the flight radius is determined; for example, if the horizontal component of wind speed is not greater than 10m / s, the flight radius is R, if the horizontal component of wind speed is greater than 10m / s and not greater than 15m / s, the flight radius is 0.8R, and if the horizontal component of wind speed is greater than 15m / s, the flight radius is 0.5R. In addition, it should be noted that if the rainfall is greater than 20mm / h, the flight radius is additionally reduced by 50%.

[0029] The way of mapping construction progress information to grid weight can be: for example, for the non-construction area: weight = 0.2 (indicating low risk); construction area: weight = 0.8 (indicating high risk); completed construction area: weight = 0.5 (indicating medium risk). Then through the

spatial join

[0030] The grid flight state determination rule is: slope ≤ 15° and weight ≤ 0.5 and within the flight radius, the flight state is normal flight; slope 15°-30° or weight 0.5-0.8, the flight state is cautious flight (such as height limit, speed reduction, etc.); slope > 30° or weight > 0.8 or beyond the flight radius, the flight state is prohibited flight. Then the

raster calculator

[0031] Reference Figure 3For S120, the specific steps of generating the original inspection path of the supervisory unmanned aerial vehicle based on the grid safety map can include S310-S340: S310, receiving node information of an inspection starting point, an inspection ending point and an inspection process node of the supervisory unmanned aerial vehicle; S320, generating all flyable paths in the grid safety map according to the node information; S330, calculating the passing costs corresponding to all flyable paths; S340, selecting a flyable path corresponding to the minimum passing cost as the original inspection path.

[0032] Specifically, the inspection node information is received in JSON format, including the starting point, the ending point and the process node, and each node includes latitude and longitude coordinates and an elevation field. The latitude and longitude are converted into plane coordinates (such as UTM projection) by an open source library (such as GeoTools) to ensure consistency with the coordinate system of the grid safety map, so that the node falls within the grid safety map. All connected paths that meet the flight conditions are traversed using a conventional path search algorithm. For example, the "no-fly" state of the grid is marked as an obstacle node by traversing the grid safety map, and only the flyable area (normal / cautious flight) is reserved as the effective search space. Each flyable grid is regarded as a graph node, and the connectivity between nodes is defined by 8-neighbor connection method (i.e. allowing the unmanned aerial vehicle to move in 8 directions up, down, left, right and diagonally), and the distance between adjacent grids is uniformly set to 1 (or weighted according to the actual distance); the process node is a must-pass point, which needs to be set as a forced passing node to limit the generation of path combinations containing these nodes in the path search. A depth-first search method is used to explore all flyable grids recursively from the starting point, record the complete path containing the starting point, process node (in order) and ending point, and then avoid repeated access to the same grid by pruning strategy, so as to obtain all flyable paths.

[0033] Reference Figure 4 For S330, the specific steps of calculating the passing costs corresponding to all flyable paths can include S410-S450: S410, obtaining the Euclidean distance between adjacent flight nodes of each flyable path; S420, obtaining the flight weight of the flyable grid covered by each flyable path; S430, obtaining the historical selection times of the inspection starting point, the inspection ending point and the inspection process node; S440, generating node weights according to the historical selection times; S450, calculating the passing costs corresponding to all flyable paths according to the Euclidean distance, the flight weight and the node weight.

[0034] The flight weight is the average weight of the path covered grid, and the average weight W f，i is the weight of the i-th grid, and n is the number of grids covered by the flyable path.

[0035] The relational database (such as MySQL) is used to store the historical selection records of the nodes, and the historical selection times of the target nodes (the starting point, the ending point, and the process nodes) in the database are queried through the SQL statement. Then, the historical selection times are converted into the node weights through the normalization processing, and the formula is: wherein, n i is the historical selection times of the node i, the denominator is the maximum selection times in all process nodes, the weight range is mapped to [0, 1], and the historical selection times of a certain node is the maximum value, so that j node = 1.

[0036] The passing cost of the flyable path j node，i is the node weight of the i-th node, and d is the total Euclidean distance of the flyable path. After calculating the passing cost corresponding to all flyable paths, the flyable path corresponding to the minimum passing cost is selected as the original inspection path.

[0037] For S130, when the supervising unmanned aerial vehicle inspects along the original inspection path, the specific way of obtaining the inspection process information can be: The unmanned aerial vehicle is equipped with a high-definition camera, an infrared thermal imager, and a laser radar, and synchronously collects the visible light image, the temperature distribution data, and the three-dimensional point cloud information of the inspection target. The defect types include track settlement, foreign object intrusion, and dangerous crack, etc. The track settlement can be calculated by the standard deviation of the sleeper point cloud height, and the standard deviation z i represents the i-th elevation value, and μ represents the mean value; if the standard deviation is greater than the set threshold value, it indicates the track settlement. The foreign object intrusion can be judged according to the point cloud density, and if the density is greater than the set threshold value, it is determined that the foreign object intrusion. The crack can be determined according to the crack pixel length and width recognized from the image and the resolution. The aspect ratio = (crack pixel length X resolution) / (crack pixel width X resolution), and when the aspect ratio is greater than the set threshold value, it is determined as the dangerous crack.

[0038] The severity of the defect can be determined according to the difference between the actual value of the defect type and the set threshold value, and the greater the difference, the more dangerous it is; the severity can be divided into three levels, high risk, medium risk, and low risk.

[0039] Referring to Figure 5 For S140, according to the defect grid labeled by the defect, the specific steps of dynamically adjusting the inspection path can include S510-S530: S510, generating the severity weight of the defect grid according to the defect labeling; S520, according to the severity weight, the defect grid is inspected and sorted; S530, according to the inspection order, the inspection starting point and the inspection end point information, the original inspection path is adjusted to generate the specified inspection path.

[0040] Specifically, the severity of each defect type corresponds to a severity weight, and the severity weight = the product of the basic weight and the influence factor coefficient; the basic weight represents the severity, for example, high risk is 1, medium risk is 0.7, and low risk is 0.3; the influence factor coefficient represents the degree of deviation, and the greater the difference, the greater the degree of deviation; the degree of deviation can also be divided into three levels, each level corresponding to an influence factor coefficient.

[0041] First, according to the severity, the initial sorting is performed, high risk - medium risk - low risk; then, according to the severity weight, the defect grid under the same severity is sorted, and the greater the severity weight, the higher the inspection priority of the defect grid.

[0042] Referring to Figure 6 For S530, according to the inspection order, the inspection starting point and the inspection end point information, the original inspection path is adjusted to generate the specified inspection path, and the specific steps can include S610-S650: S610, according to the inspection starting point and the high risk order, the first specified inspection path is generated; S620, when the inspection reaches the end point of the first specified inspection path, the first inspection starting point of the medium risk order is determined; S630, according to the first inspection starting point and the medium risk order, the second specified inspection path is generated; S640, when the inspection reaches the end point of the second specified inspection path, the second inspection starting point of the low risk order is determined; S650, according to the second inspection starting point, the low risk order and the inspection starting point, the third specified inspection path is generated.

[0043] Specifically, the inspection frequency of the first specified inspection path is greater than the inspection frequency of the second specified inspection path, which is greater than the inspection frequency of the third specified inspection path.

[0044] Taking the inspection starting point as the origin, the improved A* algorithm is used, and the "high-risk node" in the grid safety map is set as the mandatory target point to preferentially plan the shortest path through the high-risk node. The UAV GPS positioning confirms that the end point of the first designated inspection path has been reached, triggering the medium-risk path planning process; rule 1: if there is a medium-risk node within 50 meters around the high-risk path end point; rule 2: if there is no adjacent medium-risk node, the high-risk path end point is defaulted as the starting point of the medium-risk inspection; the greedy method is used to plan the path: starting from the first inspection starting point, the nearest un-inspected medium-risk node is selected each time until all medium-risk nodes are covered. After reaching the medium-risk path end point, the low-risk node set is read, and the spatial distribution density is counted. If the low-risk nodes are concentrated around the medium-risk end point (e.g. 80% of the nodes are within a 200-meter range), the medium-risk end point is taken as the second inspection starting point; otherwise, the geometric center of the low-risk node cluster is selected as the starting point. On the basis of covering all low-risk nodes, the final return to the initial inspection starting point is ensured, forming a "closed loop path".

[0045] Further, when the supervising UAV inspects along the designated inspection path, the percentage of the power of the supervising UAV needs to be obtained; then when the percentage of the power is lower than the set inspection power value, the return inspection path is generated according to the current percentage of the power, the inspection progress and the set minimum inspection power.

[0046] Specifically, real-time voltage and current data can be obtained by the unmanned aerial vehicle power management unit to obtain the percentage of power, and when the real-time percentage of power is less than or equal to a set inspection power value (such as 30%), a return path generation process is started. A minimum inspection power (such as 15%) is set as a forced return threshold, and if the power is lower than this value, the inspection is immediately interrupted and the unmanned aerial vehicle returns. In the unmanned aerial vehicle task planning system, a unique identifier (such as Route_1 / Route_2 / Route_3) is assigned to the first / second / third designated inspection path. The real-time coordinates of the unmanned aerial vehicle are obtained by GPS positioning and matched with the node list of the preset path: if the real-time coordinates are within the node sequence range of Route_1, it is determined that the first designated inspection path stage is reached. Similarly, the current inspection path stage is determined. The difference between the current power and the minimum inspection power is calculated: the remaining available power = current power - minimum inspection power. The estimated remaining inspection power is estimated based on historical energy consumption data: estimated remaining energy consumption = remaining node number x node average energy consumption. If the remaining available power is greater than or equal to the estimated remaining energy consumption, the inspection continues; otherwise, the unmanned aerial vehicle returns. When the inspection continues, the shortest energy consumption path is searched as the return inspection path in the grid safety map, with the current location as the starting point and the take-off point as the end point, and the un-inspected nodes are marked for direct inspection after charging is completed. It should be noted that if the current location of the supervisory unmanned aerial vehicle is in the first designated inspection path, the shortest energy consumption path needs to include the medium-risk nodes and low-risk nodes; if the current location of the supervisory unmanned aerial vehicle is in the second designated inspection path, the shortest energy consumption path needs to include the low-risk nodes.

[0047] The implementation principle of the embodiment is: The terrain data, weather data and construction progress information of the target inspection area are received, and then a blank grid safety map of the target inspection area is generated according to the terrain data. Then the slope data in the terrain data is extracted, and the flight radius is determined according to the weather data. Then the construction progress information is mapped into a grid weight, and each grid in the blank grid safety map is given a corresponding flight state according to the slope data, the flight radius and the grid weight, to generate a final grid safety map. Then the node information of the inspection starting point, the inspection end point and the inspection process nodes of the supervisory unmanned aerial vehicle is received, and all flyable paths are generated in the grid safety map according to the node information. The Euclidean distance between adjacent flight nodes of each flyable path is obtained, and the flight weight of the grid covered by each flyable path is obtained. The node weight is generated according to the historical selection times of the inspection starting point, the inspection end point and the inspection process nodes. Then the passing cost corresponding to all flyable paths is calculated according to the Euclidean distance, the flight weight and the node weight, and the flyable path corresponding to the minimum passing cost is selected as the original inspection path.

[0048] When the supervisory unmanned aerial vehicle patrols along the original patrol path, patrol process information is acquired, and then defect labeling is performed on the grid safety map according to the patrol process information and the flight position of the supervisory unmanned aerial vehicle; according to the defect labeling, a defect grid severity weight is generated, and then the defect grid is sorted according to the severity weight, a first specified patrol path is generated according to the patrol starting point and the high-risk sorting, and when the patrol reaches the end point of the first specified patrol path, a first patrol starting point of the medium-risk sorting is determined; then a second specified patrol path is generated according to the first patrol starting point and the medium-risk sorting; and when the patrol reaches the end point of the second specified patrol path, a second patrol starting point of the low-risk sorting is determined, and then a third specified patrol path is generated according to the second patrol starting point, the low-risk sorting and the patrol starting point.

[0049] In addition, when the supervisory unmanned aerial vehicle patrols along the specified patrol path, the percentage of the power of the supervisory unmanned aerial vehicle needs to be acquired; and then when the percentage of the power is lower than the set patrol power value, a return patrol path is generated according to the current percentage of the power, the patrol progress and the set minimum patrol power.

[0050] Based on the above method embodiments, the second embodiment of the present application discloses a path planning and defect labeling system for supervisory unmanned aerial vehicle patrol. The path planning and defect labeling system for supervisory unmanned aerial vehicle patrol of the embodiment of the present application can implement any one of the above-mentioned path planning and defect labeling methods for supervisory unmanned aerial vehicle patrol, and the specific working process of each module in the path planning and defect labeling system for supervisory unmanned aerial vehicle patrol can refer to the corresponding process in the above method embodiments.

[0051] For ease of understanding, an example is as follows: a path planning and defect labeling system for supervisory unmanned aerial vehicle patrol, comprising: a map acquisition module, configured to acquire a grid safety map of a target patrol area; a path generation module, configured to generate an original patrol path for supervisory unmanned aerial vehicle patrol based on the grid safety map; an information acquisition module, configured to acquire patrol process information when the supervisory unmanned aerial vehicle patrols along the original patrol path; a defect labeling module, configured to perform defect labeling on the grid safety map according to the patrol process information and the flight position of the supervisory unmanned aerial vehicle; a path adjustment module, configured to dynamically adjust the patrol path according to the grid corresponding to the defect labeling.

[0052] The above are preferred embodiments of the present application, and do not limit the protection scope of the present application in sequence, any feature disclosed in the specification (including the abstract and the drawings) can be replaced by other equivalent or similar features unless specifically described. That is, each feature is only an example of a series of equivalent or similar features unless specifically described.

Claims

1. A method for path planning and defect marking for inspection by a supervisory drone, characterized in that: include: Obtain a grid security map of the target inspection area; Based on the grid safety map, an original inspection path for the supervisory drone inspection is generated; When the supervisory drone inspects along the original inspection route, obtaining inspection process information; Marking defects on the grid safety map based on the inspection process information and the flight position of the supervisory drone; The inspection path is dynamically adjusted according to the defect grid marked with the defect.

2. A method for path planning and defect marking for inspection by a supervisory drone according to claim 1, characterized in that: The step of obtaining a grid safety map of the target inspection area includes: Receive terrain data, meteorological data and construction progress information of the target inspection area; Generating a blank grid safety map of the target inspection area based on the terrain data; extracting slope data from the terrain data; determining a flight radius based on the meteorological data; Mapping the construction progress information into grid weights; According to the slope data, the flight radius and the grid weight, a corresponding flight state is assigned to each grid in the blank grid safety map to generate a final grid safety map.

3. The method for path planning and defect marking for inspection by a supervisory drone according to claim 1, characterized in that: The step of generating an original inspection path for the inspection by the supervisory drone based on the grid safety map includes: Receive node information of the inspection starting point, inspection end point and inspection process nodes of the supervision drone; generating all flyable paths in the grid safety map according to the node information; Calculate the travel costs corresponding to all flyable paths; The flyable path corresponding to the minimum travel cost is selected as the original inspection path.

4. A method for path planning and defect marking for inspection by a supervisory drone according to claim 3, characterized in that: The step of calculating the travel costs corresponding to all flyable paths includes: Get the Euclidean distance between adjacent flight nodes of each flyable path; Get the flight weight of each flyable path corresponding to the grid; Obtain the historical selection times of the inspection starting point, the inspection end point and the inspection process node; Generate node weights based on the number of historical selections; The travel costs corresponding to all flyable paths are calculated according to the Euclidean distance, the flight weight, and the node weight.

5. The method for path planning and defect marking for inspection by a supervisory drone according to claim 1, characterized in that: The step of dynamically adjusting the inspection path according to the defect grid marked with the defect comprises: generating a severity weight of a defect grid according to the defect annotation; According to the severity weight, the defective grids are inspected and sorted; The original inspection path is adjusted according to the inspection sequence, the inspection starting point, and the inspection end point information to generate a designated inspection path.

6. A method for path planning and defect marking for inspection by a supervisory UAV according to claim 5, characterized in that: The step of adjusting the original inspection path according to the inspection sequence, the inspection starting point, and the inspection end point information to generate a designated inspection path includes: The inspection ranking includes high risk, medium risk and low risk; Generate a first designated inspection path according to the inspection starting point and high-risk ranking; When the inspection reaches the end point of the first designated inspection path, determining the first inspection starting point of the medium-risk ranking; Generate a second designated inspection path based on the first inspection starting point and the medium-risk ranking; When the inspection reaches the end point of the second designated inspection path, determining a second inspection starting point with a low risk ranking; generating a third designated inspection path according to the second inspection starting point, the low-risk ranking, and the inspection starting point; The number of inspections of the first designated inspection path is greater than the number of inspections of the second designated inspection path, which is greater than the number of inspections of the third designated inspection path.

7. A method for path planning and defect marking for inspection by a supervisory UAV according to claim 6, characterized in that: The inspection path planning and defect marking method further includes: When the supervisory drone is inspecting along the designated inspection route, obtaining a battery percentage of the supervisory drone; When the power percentage is lower than the set inspection power value, a return inspection route is generated according to the current power percentage, the inspection progress and the set minimum inspection power.

8. A supervisory drone inspection path planning and defect marking system, characterized in that: Executing the method for path planning and defect marking for inspection by a supervisory drone as described in any one of claims 1 to 7, comprising: A map acquisition module is used to obtain a grid safety map of the target inspection area; A path generation module is used to generate an original inspection path for the supervisory drone inspection based on the grid safety map; An information acquisition module, configured to acquire inspection process information when the supervisory drone inspects along the original inspection path; A defect marking module, configured to mark defects on the grid safety map based on the inspection process information and the flight position of the supervisory drone; The path adjustment module is used to dynamically adjust the inspection path according to the defect grid marked with the defect.

9. A terminal, characterized in that: include: Memory, which stores the path planning and defect marking programs used by the supervisory drone inspection; A processor is used to execute the program stored on the memory to implement the steps of the path planning and defect marking method for inspection by a supervisory drone as described in any one of claims 1 to 7.

Images