Unmanned aerial vehicle indoor inspection method based on laser radar positioning

Abstract

The invention relates to an unmanned aerial vehicle indoor inspection method based on laser radar positioning, and belongs to the technical field of unmanned aerial vehicle control. The method comprises the following steps: acquiring indoor environment information through a laser radar to obtain indoor point cloud data, performing angular point detection on the indoor point cloud data, extracting and identifying angular points, removing redundant angular points to obtain real-time point cloud data, and constructing an incremental map according to the real-time point cloud data and an indoor actual environment position; position observation data of the unmanned aerial vehicle is obtained through a mapping relation between real-time point cloud data and a map in an incremental map, a routing inspection path of the unmanned aerial vehicle is constructed according to the incremental map and initial pose information of the unmanned aerial vehicle, the routing inspection path is optimized and guided through the position observation data, and after a routing inspection task is completed, the unmanned aerial vehicle automatically returns to enter a cabin and is automatically charged. Indoor obstacle avoidance and accurate positioning of automatic inspection of the unmanned aerial vehicle are realized, and stable operation of the unmanned aerial vehicle is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle control, and in particular relates to an indoor inspection method of an unmanned aerial vehicle based on laser radar positioning. Background Art

[0002] As an aerial robot, drones can carry a variety of sensor devices during the execution of tasks due to their unique technical characteristics. They can monitor the current environment in real time, determine their own location, adjust their flight attitude in real time, and avoid various obstacles, so as to complete the given tasks well. During the flight of drones, due to their unique characteristics and uncertain environmental factors, drones are particularly susceptible to the influence of their own performance and terrain. Specifically, when flying outdoors, drones usually rely on global positioning system (GPS) signals to determine their own positions. At least 3 to 4 GPS signals are required for accurate positioning. In indoor environments, GPS signals often cannot penetrate buildings, resulting in drones being unable to rely on traditional positioning methods. Indoor environments are complex and there are a large number of obstacles. Drones need to have good obstacle avoidance capabilities to avoid collisions and ensure flight safety. Indoor space is usually limited, so drones need to have good endurance and flexible energy supply solutions. In indoor environments, drones lack the ability to make autonomous decisions and need to make corresponding flight actions and path planning according to environmental changes and mission requirements. Summary of the invention

[0003] In order to solve the above problems existing in the prior art, the present invention provides a UAV indoor inspection method based on laser radar positioning;

[0004] The purpose of the present invention can be achieved through the following technical solutions:

[0005] A drone indoor inspection method based on laser radar positioning, comprising:

[0006] Collecting indoor environmental information through a laser radar to obtain indoor point cloud data, performing corner detection and extraction on the indoor point cloud data and identifying corner points, removing redundant corner points to obtain real-time point cloud data, and constructing an incremental map based on the real-time point cloud data and the actual indoor environmental position;

[0007] The position observation data of the UAV is obtained through the mapping relationship between the real-time point cloud data and the map in the incremental map, and the inspection path of the UAV is constructed according to the incremental map and the initial posture information of the UAV. The inspection path is optimized and guided by the position observation data. After the inspection task is completed, the UAV automatically returns to the cabin and charges automatically.

[0008] Specifically, the corner point detection obtains the shape feature corner points in the local area by calculating the point cloud features based on the local curvature and normal vector information of the point cloud data; the normal vector is obtained by performing principal component analysis on the point coordinates in the nearest neighbor point set of the point cloud data, and the eigenvector corresponding to the minimum eigenvalue is obtained, which is the normal vector; the curvature of each point in the local area is calculated using the normal vector.

[0009] Specifically, the incremental map is dynamically updated based on the incremental data packets generated by the real-time map features parsed from the real-time point cloud data and the real-time UAV posture information; the incremental map replaces the corresponding map space according to the incremental data packets.

[0010] Specifically, the inspection path construction method is as follows: constructing the UAV flight environment constraints and flight performance constraints through the incremental map; the flight performance constraints include the maximum flight distance, the minimum flight distance and the climbing angle of the UAV; the flight environment constraints are to mark indoor obstacles as threat obstacle sources, and construct the flight environment constraints by calculating the distance between the UAV and each threat source;

[0011] Set the path starting point and end point, establish a path network according to the incremental map, mark the threat obstacle source in the path network through the flight environment constraint, take the path starting point as the initial node, perform nearest neighbor search through local sensitive hashing, use the search result as the feasible path node, obtain a feasible path set by iteratively searching the path network, establish a path cost function according to the feasible path node, and calculate it as:

[0012]

[0013] Where L(path) is the path cost, N is the total number of nodes, i is the number of feasible path nodes, and R i is the path node gain, k is the node path gain, F m is the flight performance constraint, s max is the maximum flight distance, s min is the minimum flight distance, a is the climbing angle, Q m For flight environment constraints;

[0014] The node corresponding to the minimum cost is obtained by calculating the path cost function, and a feasible path is extracted from the feasible path set according to the node.

[0015] Specifically, the sensitivity characteristics of the local sensitive hashing are set by the flight environment constraints.

[0016] Specifically, the optimization guidance method is to update the path network through the position observation data detected in real time, and optimize the connection paths between the nodes of the path network. The calculation formula is:

[0017]

[0018] Among them, p(u) is the node set of a single feasible path, i is the node count, n is the total number of nodes on the path, and p i is the i-th control node, B ik (u) is the locally updated planning connection path basis function;

[0019] The path cost function is reconstructed according to the updated path network information, and the calculation formula is:

[0020] Ltcn=wLcn+(1-w)ct m ,

[0021] Among them, L tcn is the updated path cost function, w is the node update parameter, L cn is the path cost function before updating, c is the time reward coefficient, t m Plan the remaining running time for the route.

[0022] Specifically, the incremental map sets target points in the indoor plan and performs indoor plan and elevation line planning based on the indoor UAV integrated automatic inspection platform, and combines the unmanned automatic hangar to automatically launch, store and land the UAV.

[0023] Specifically, the unmanned automatic hangar is connected to the hangar's control system through contact technology, and the system automatically identifies the drone and provides it with take-off instructions. After completing the mission, the drone will automatically return to the hangar and connect to the hangar's control system through contact technology. The system automatically identifies the drone and provides it with landing instructions. During the landing process, the hangar's landing guidance system will ensure that the drone lands safely at the designated location.

[0024] Specifically, the unmanned automatic hangar and indoor UAV integrated automatic inspection platform supports remote control and multi-terminal operation. The landing guidance method is to install independent embedded vision systems on both sides of the runway of the unmanned automatic hangar. The embedded vision system sends high-level control commands to the UAV, records GNSS data and estimated data, and transmits image processing results and PTU status to the navigation computer, which calculates the estimated relative position of the UAV; because the ground station determines to switch from GNSS auxiliary mode to visual auxiliary mode, the autopilot guides the UAV according to the correct data received by the XTend modem.

[0025] The beneficial effects of the present invention are:

[0026] By using the observed map features to locate the drone's own position and posture during its movement, and building an incremental map based on this, the drone can understand its own position in real time and create a detailed indoor map to provide a basis for subsequent navigation and task execution; the position of indoor objects is determined through point cloud data collected by the lidar, and the location of obstacle sources in the path planning process is further determined; by continuously comparing the point cloud data collected at different time points, the relative movement distance and posture changes of the lidar can be calculated, the positioning of the drone's own position is completed, and the map information is updated synchronously, realizing a closed-loop incremental update of the dynamic environment map; local path planning is carried out according to point cloud data, getting rid of the dependence on precise environmental models, and being able to adapt to various unknown and dynamic environmental conditions more flexibly; by optimizing and guiding the inspection path, the drone can quickly adapt to environmental changes in a shorter re-planning time. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0028] Figure 1 The present invention is a flowchart of a method for indoor inspection using a drone based on laser radar positioning. DETAILED DESCRIPTION

[0029] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0030] See also Figure 1 ,A UAV indoor inspection method based on lidar positioning;

[0031] A drone indoor inspection method based on laser radar positioning, comprising:

[0032] Collecting indoor environmental information through a laser radar to obtain indoor point cloud data, performing corner detection and extraction on the indoor point cloud data and identifying corner points, removing redundant corner points to obtain real-time point cloud data, and constructing an incremental map based on the real-time point cloud data and the actual indoor environmental position;

[0033] The position observation data of the UAV is obtained through the mapping relationship between the real-time point cloud data and the map in the incremental map, and the inspection path of the UAV is constructed according to the incremental map and the initial posture information of the UAV. The inspection path is optimized and guided by the position observation data. After the inspection task is completed, the UAV automatically returns to the cabin and charges automatically.

[0034] In this embodiment, the SLAM (Simultaneous Localization And Mapping) algorithm is used to determine the position of an object by emitting laser pulses and measuring the time required for the reflected light to obtain information about the surrounding environment, and to build a map of the indoor environment in real time based on the environmental information; the ISS corner detection algorithm for point cloud data is based on the local curvature and normal vector information of the point cloud data, and identifies points with significant curvature changes by calculating point cloud features.

[0035] Specifically, the corner point detection obtains the shape feature corner points in the local area by calculating the point cloud features based on the local curvature and normal vector information of the point cloud data; the normal vector is obtained by performing principal component analysis on the point coordinates in the nearest neighbor point set of the point cloud data, and the eigenvector corresponding to the minimum eigenvalue is obtained, which is the normal vector; the curvature of each point in the local area is calculated using the normal vector.

[0036] Specifically, the incremental map is dynamically updated based on the incremental data packets generated by the real-time map features parsed from the real-time point cloud data and the real-time UAV posture information; the incremental map replaces the corresponding map space according to the incremental data packets.

[0037] In this embodiment, the incremental data packet is an incremental packet of a general format generated by the dynamic update of the temporal database navigation information, which is then converted into a custom format (binary format data file) and then generated by the server side, and handed over to the information publishing system of the server side to transmit the incremental packet to the control terminal through data transmission means, thereby performing an incremental update of the incremental map data. Specifically, when performing the incremental update operation, for each map sheet, the incremental data is written into the reserved space of the map sheet, and the disk space is regularly sorted and rewritten to rewrite all the data.

[0038] Specifically, the inspection path construction method is as follows: constructing the UAV flight environment constraints and flight performance constraints through the incremental map; the flight performance constraints include the maximum flight distance, the minimum flight distance and the climbing angle of the UAV; the flight environment constraints are to mark indoor obstacles as threat obstacle sources, and construct the flight environment constraints by calculating the distance between the UAV and each threat source;

[0039] Set the path starting point and end point, establish a path network according to the incremental map, mark the threat obstacle source in the path network through the flight environment constraint, take the path starting point as the initial node, perform nearest neighbor search through local sensitive hashing, use the search result as the feasible path node, obtain a feasible path set by iteratively searching the path network, establish a path cost function according to the feasible path node, and calculate it as:

[0040]

[0041] Where L(path) is the path cost, N is the total number of nodes, i is the number of feasible path nodes, and R i is the path node gain, k is the node path gain, F m is the flight performance constraint, s max is the maximum flight distance, s min is the minimum flight distance, a is the climbing angle, Q m Due to flight environment constraints;

[0042] The node corresponding to the minimum cost is obtained by calculating the path cost function, and a feasible path is extracted from the feasible path set according to the node.

[0043] Specifically, the sensitivity characteristics of the local sensitive hashing are set by the flight environment constraints.

[0044] Specifically, the optimization guidance method is to update the path network through the position observation data detected in real time, and optimize the connection paths between the nodes of the path network. The calculation formula is:

[0045]

[0046] Among them, p(u) is the node set of a single feasible path, i is the node count, n is the total number of nodes on the path, and p i is the i-th control node, B ik (u) is the locally updated planning connection path basis function;

[0047] The path cost function is reconstructed according to the updated path network information, and the calculation formula is:

[0048] L tcn =wL cn +(1-w)ct m ,

[0049] Among them, L tcn is the updated path cost function, w is the node update parameter, L cn is the path cost function before updating, c is the time reward coefficient, t m Plan the remaining running time for the route.

[0050] Specifically, the incremental map sets target points in the indoor plan and performs indoor plan and elevation line planning based on the indoor UAV integrated automatic inspection platform, and combines the unmanned automatic hangar to automatically launch, store and land the UAV.

[0051] Specifically, the unmanned automatic hangar is connected to the hangar's control system through contact technology, and the system automatically identifies the drone and provides it with take-off instructions. After completing the mission, the drone will automatically return to the hangar and connect to the hangar's control system through contact technology. The system automatically identifies the drone and provides it with landing instructions. During the landing process, the hangar's landing guidance system will ensure that the drone lands safely at the designated location.

[0052] Specifically, the unmanned automatic hangar and indoor UAV integrated automatic inspection platform supports remote control and multi-terminal operation. The landing guidance method is to install independent embedded vision systems on both sides of the runway of the unmanned automatic hangar. The embedded vision system sends high-level control commands to the UAV, records GNSS data and estimated data, and transmits image processing results and PTU status to the navigation computer, which calculates the estimated relative position of the UAV; because the ground station determines to switch from GNSS auxiliary mode to visual auxiliary mode, the autopilot guides the UAV according to the correct data received by the XTend modem.

[0053] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A drone indoor inspection method based on laser radar positioning, characterized in that: include: Collecting indoor environmental information through laser radar to obtain indoor point cloud data, performing corner detection and extraction on the indoor point cloud data and identifying corner points, removing redundant corner points to obtain real-time point cloud data, and constructing an incremental map based on the real-time point cloud data and the actual indoor environmental position; The position observation data of the UAV is obtained through the mapping relationship between the real-time point cloud data and the map in the incremental map, and the inspection path of the UAV is constructed according to the incremental map and the initial posture information of the UAV. The inspection path is optimized and guided by the position observation data. After the inspection task is completed, the UAV automatically returns to the cabin and charges automatically.

2. The indoor inspection method of a drone based on laser radar positioning according to claim 1 is characterized in that: The corner point detection obtains the shape feature corner points in the local area by calculating the point cloud features based on the local curvature and normal vector information of the point cloud data; the normal vector is obtained by performing principal component analysis on the point coordinates in the nearest neighbor point set of the point cloud data, and the eigenvector corresponding to the minimum eigenvalue is obtained, which is the normal vector; the curvature of each point in the local area is calculated using the normal vector.

3. The indoor inspection method of a drone based on laser radar positioning according to claim 1 is characterized in that: The incremental map is dynamically updated based on the incremental data packets generated by the real-time map features parsed from the real-time point cloud data and the real-time UAV posture information; the incremental map replaces the corresponding map space according to the incremental data packets.

4. The indoor inspection method of a drone based on laser radar positioning according to claim 1 is characterized in that: The inspection path construction method is as follows: constructing the UAV flight environment constraints and flight performance constraints through the incremental map; the flight performance constraints include the maximum flight distance, the minimum flight distance and the climbing angle of the UAV; the flight environment constraints are to mark indoor obstacles as threat obstacle sources, and construct the flight environment constraints by calculating the distance between the UAV and each threat source; Set the path starting point and end point, establish a path network according to the incremental map, mark the threat obstacle source in the path network through the flight environment constraint, take the path starting point as the initial node, perform nearest neighbor search through local sensitive hashing, use the search result as the feasible path node, obtain a feasible path set by iteratively searching the path network, establish a path cost function according to the feasible path node, and calculate it as: Where L(path) is the path cost, N is the total number of nodes, i is the number of feasible path nodes, and R i is the path node gain, k is the node path gain, F m is the flight performance constraint, s max is the maximum flight distance, s min is the minimum flight distance, a is the climbing angle, Q m Due to flight environment constraints; The node corresponding to the minimum cost is obtained by calculating the path cost function, and a feasible path is extracted from the feasible path set according to the node.

5. The indoor inspection method of a drone based on laser radar positioning according to claim 4 is characterized in that: The sensitivity characteristics of the local sensitive hashing are set by the flight environment constraints.

6. The indoor inspection method of a drone based on laser radar positioning according to claim 1 is characterized in that: The optimization guidance method is to update the path network through the position observation data detected in real time, and optimize the connection paths between the nodes of the path network. The calculation formula is: Among them, p(u) is the node set of a single feasible path, i is the node count, n is the total number of nodes on the path, and p i is the i-th control node, B ik (u) is the locally updated planning connection path basis function; The path cost function is reconstructed according to the updated path network information, and the calculation formula is: L tcn =wL cn +(1-w)ct m , Among them, L tcn is the updated path cost function, w is the node update parameter, L cn is the path cost function before updating, c is the time reward coefficient, t m Plan the remaining running time for the route.

7. The indoor inspection method of a drone based on laser radar positioning according to claim 1 is characterized in that: The incremental map sets target points in the indoor plan and performs indoor plan and elevation line planning based on the indoor UAV integrated automatic inspection platform, and combines the unmanned automatic hangar to automatically launch, store and land the UAV.

8. The indoor inspection method of a drone based on laser radar positioning according to claim 7 is characterized in that: The unmanned automatic hangar is connected to the hangar's control system through contact technology. The system automatically identifies the drone and provides it with take-off instructions. After completing the mission, the drone will automatically return to the hangar and connect to the hangar's control system through contact technology. The system automatically identifies the drone and provides it with landing instructions. During the landing process, the hangar's landing guidance system will ensure that the drone lands safely at the designated location.

9. The indoor inspection method of a drone based on laser radar positioning according to claim 7 is characterized in that: The unmanned automatic hangar and indoor UAV integrated automatic inspection platform supports remote control and multi-terminal operation. The landing guidance method is to install independent embedded vision systems on both sides of the runway of the unmanned automatic hangar. The embedded vision system sends high-level control commands to the UAV, records GNSS data and estimated data, and transmits image processing results and PTU status to the navigation computer, which calculates the estimated relative position of the UAV; because the ground station determines to switch from the GNSS auxiliary mode to the visual auxiliary mode, the autopilot guides the UAV according to the correct data received by the XTend modem.

Images