Underground mine environment mapping positioning method and system
Through the tightly coupled fusion of laser inertial navigation and SC descriptor matching technology, the problem of high positioning drift rate in underground mining environments has been solved, and efficient and reliable mapping and positioning have been achieved, meeting the precision requirements of unmanned loaders.
Patent Information
- Application Number
- CN202510725428.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-30
- Publication Date
- 2025-10-03
AI Technical Summary
In underground mines, where there is no GPS signal and lighting conditions are poor, traditional mapping and positioning algorithms have high positioning drift rates and insufficient reliability under low-feature structures, making it difficult to meet the precision requirements of unmanned loaders.
The global point cloud map of the underground mine environment is constructed by the tightly coupled fusion of laser inertial navigation. The SC descriptors in the SC database are used for point cloud matching and pose correction to eliminate noise points. The positioning process is optimized through the SLAM algorithm and Euclidean distance clustering algorithm.
The positioning reliability and mapping accuracy of unmanned loaders in underground mine environments are improved, and efficient and robust relocalization is achieved under low-feature structures.
Smart Images

Figure CN120740563A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method and system for mapping and positioning an underground mine environment, belonging to the technical field of mapping and positioning. Background Art
[0002] As underground mining transitions toward intelligent and unmanned operations, equipment such as underground loaders (LHDs) are in urgent need of mapping and positioning technology in mine environments. Underground mines lack GPS signals, experience poor lighting conditions, and are prone to low-feature structures such as long, straight tunnels and smooth rock walls. Interference from dust and water mist easily generates a large amount of noise in laser point clouds. Traditional mapping and positioning algorithms, which rely on feature extraction, suffer from high positioning drift and insufficient mapping reliability in mine conditions, making them unable to meet the precision requirements for remote control and autonomous operation of LHDs. Summary of the Invention
[0003] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a method and system for mapping and positioning an underground mine environment, overcome the interference of laser point cloud noise, improve the positioning reliability and mapping accuracy of unmanned scrapers, and meet the needs of intelligent and unmanned operation.
[0004] To achieve the above object, the present invention is implemented by adopting the following technical solutions: In one aspect, the present invention provides a method for mapping and positioning an underground mine environment, comprising: Constructing a global point cloud map of the underground mine environment and an SC database, wherein the SC database includes SC descriptors of multiple frames of point cloud data; Obtain real-time underground mine environment point cloud data; Matching the current frame underground mine environment point cloud data with the global point cloud map of the underground mine environment to obtain the pose and matching score of the current frame point cloud data in the global point cloud map of the underground mine environment; Determine whether the scraper positioning is drifting based on the pose and matching score of the current frame point cloud data in the global point cloud map of the underground mine environment: If the positioning of the scraper does not drift, the position of the current frame point cloud data in the global point cloud map of the underground mine environment is used as the positioning result; If the scraper drifts in positioning, the SC descriptor of the current frame point cloud data is generated, and the SC descriptor of the current frame point cloud data is matched with the SC descriptor of the frame point cloud data in the SC database. If the match is successful, the posture is corrected and used as the positioning result. If the match fails, it is determined to be positioning drift.
[0005] Furthermore, the method for constructing a global point cloud map of the underground mine environment includes: Obtain underground mine environment point cloud data and scraper angular velocity and linear acceleration data; The underground mine environment point cloud data and the angular velocity and linear acceleration data of the scraper are used as input, and the global point cloud map of the underground mine environment is obtained based on the SLAM algorithm output.
[0006] Furthermore, the method for constructing the SC database includes: Build SC database; Acquire point cloud data of underground mine environment; Generate SC descriptors based on each frame of underground mine environment point cloud data; It is determined whether each frame of underground mine environment point cloud data is a key frame. If it is a key frame, the SC descriptor of the frame of underground mine environment point cloud data is stored in the SC database.
[0007] Furthermore, generating the SC descriptor according to each frame of underground mine environment point cloud data includes: A frame of underground mine environment point cloud data is divided into a plurality of sectors according to azimuth angles, and is divided into a plurality of rings according to radius, and the intersection of each sector and ring is used as a block area; The maximum height of the point cloud in each block is used as the identification number of the block. For blocks without point clouds, the identification number of the block is set to 0. The identification numbers of all blocks are combined into a two-dimensional matrix, and the two-dimensional matrix is used as the SC descriptor of the underground mine environment point cloud data of this frame; The azimuth angle and radius are determined by the laser radar coordinate system.
[0008] Furthermore, each frame of the underground mine environment point cloud data is judged as a key frame only if it satisfies at least one of the following conditions: The moving distance of the scraper in this frame of point cloud data exceeds the preset distance; The yaw angle of the scraper in this frame of point cloud data exceeds the set angle; The cosine similarity between the SC descriptor of the frame and the nearest adjacent key frame is lower than a preset threshold; The change in the proportion of invalid sectors in the point cloud data of this frame exceeds a preset threshold.
[0009] Furthermore, after obtaining the real-time underground mine environment point cloud data and the underground mine environment point cloud data, pre-processing is also performed on the data. The pre-processing includes: The underground mine point cloud data is processed using voxel grid filtering to reduce the point cloud density and obtain the downsampled underground mine point cloud data. The downsampled underground mine point cloud data is filtered using the Euclidean distance clustering method to obtain preprocessed real-time underground mine environment point cloud data and underground mine environment point cloud data.
[0010] Furthermore, judging whether the positioning of the scraper is drifting based on the pose of the current frame point cloud data in the global point cloud map of the underground mine environment and the matching score includes: Compare the matching score between the current frame point cloud data and the global point cloud map of the underground mine environment with a preset threshold. If the matching score is less than the preset threshold, it is determined that the scraper positioning is drifting. If the matching score is greater than a preset threshold, the pose distance between the pose of the current frame point cloud data in the global point cloud map of the underground mine environment and the pose of the previous frame point cloud data in the global point cloud map of the underground mine environment is calculated; When the pose distance between the pose of the current frame point cloud data in the global point cloud map of the underground mine environment and the pose of the previous frame point cloud data in the global point cloud map of the underground mine environment meets the following constraint conditions, it is determined that the scraper is drifting in positioning; otherwise, it is determined that the scraper is not drifting in positioning. The expression of the constraint conditions is: ; in, represents the pose distance, Indicates time interval The average speed within Indicates the interval between obtaining the current frame of point cloud data and obtaining the previous frame of point cloud data.
[0011] Furthermore, matching the SC descriptor of the current frame point cloud data with the SC database includes: S1. averaging the values of each row of the matrix of the SC descriptors of each frame of point cloud data in the SC database and the current frame of point cloud data to obtain a one-dimensional vector, and using the one-dimensional vector as the ring key; S2. Traverse the ring key of each frame point cloud data in the SC database and calculate the Euclidean distance between it and the ring key of the current frame point cloud data; S3, select the SC descriptor of the 5 frames of point cloud data with the smallest Euclidean distance; S4, rotating the SC descriptor of each frame point cloud data, and calculating the cosine similarity between the SC descriptor of the rotated frame point cloud data and the SC descriptor of the current frame; S5. Compare the cosine similarity of the SC descriptors of the five rotated frame point cloud data with the set threshold. If any cosine similarity is greater than the set threshold, the match is successful. Otherwise, continue S4 to S5 until the rotation angle reaches 360°. If the cosine similarity is still less than the set threshold, the match fails. The rotation angle corresponding to the successful matching is used as the yaw angle deviation, and the yaw angle deviation is used for posture correction.
[0012] Furthermore, the number of rotations is determined by the number of sectors of the SC descriptor of the frame point cloud data, and the angle of each rotation is the same and the sum of the angles is 360°.
[0013] On the other hand, the present invention also provides an underground mine environment mapping and positioning system for implementing any of the above-mentioned underground mine environment mapping and positioning methods, which includes a laser radar, an inertial measurement unit, and a mapping and positioning system, wherein the laser radar is used to collect underground mine environment point cloud data, and the inertial measurement unit is used to obtain the position information of the scraper; The laser radar is arranged above the inertial measurement unit, and the inertial measurement unit is arranged above the cab of the scraper through a bracket. The laser radar and the inertial measurement unit are both electrically connected to the mapping and positioning system. The mapping and positioning system includes a mapping module, an SC database construction module, a positioning module and a repositioning module. The mapping module is configured to construct a global point cloud map of the underground mine environment; The SC database construction module is configured to construct the SC database; The positioning module is configured to obtain real-time underground mine environment point cloud data; match the current frame underground mine environment point cloud data with the global point cloud map of the underground mine environment to obtain the position and posture of the current frame point cloud data in the global point cloud map of the underground mine environment; and determine whether the positioning of the scraper is drifting based on the position and posture of the current frame point cloud data in the global point cloud map of the underground mine environment; The repositioning module is configured to generate the SC descriptor of the current frame point cloud data when the scraper drifts in positioning, and match the SC descriptor of the current frame point cloud data with the SC descriptor of the frame point cloud data in the SC database. If the match is successful, the posture is corrected and used as the positioning result. If the match fails, it is determined to be positioning drift.
[0014] Compared with the prior art, the present invention has the following beneficial effects: This method addresses the low-feature structure characteristics of underground mines, such as the lack of GPS signals, poor lighting conditions, and long, straight tunnels and smooth rock walls. It uses a tightly coupled laser inertial navigation fusion approach to achieve mapping and positioning. The original point cloud is directly registered to the map, fully utilizing the tiny features in the environment. A compact scene descriptor is generated during the mapping phase, enabling efficient and robust relocalization when positioning drift occurs. The present invention adopts a clustering method to eliminate noise points, solving the problem of ineffective filtering of traditional Gaussian filtering, half-wave filtering and the like under the low-feature structure characteristics of underground mines. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 A schematic diagram of the mapping phase of an underground mine environment mapping and positioning method according to an embodiment of the present invention; Figure 2 A schematic diagram of the positioning phase of an underground mine environment mapping and positioning method according to an embodiment of the present invention; Figure 3 Schematic diagram of the process of the Euclidean distance clustering algorithm of the underground mine environment mapping and positioning method in one embodiment of the present invention; Figure 4 A schematic diagram of the installation of a laser radar and an inertial measurement unit for an underground mine environment mapping and positioning method according to an embodiment of the present invention; Figure 5 Schematic diagram of SC descriptor calculation in an underground mine environment mapping and positioning method according to an embodiment of the present invention; In the figure: 1-LiDAR, 2-Inertial Measurement Unit. DETAILED DESCRIPTION
[0016] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. Example 1
[0017] like Figure 1 As shown, an embodiment of the present invention provides a method for mapping and positioning an underground mine environment. The overall method is divided into a mapping stage and a positioning stage.
[0018] The process of the mapping stage is as follows Figure 1 As shown, the specific steps include: Obtain underground mine environment point cloud data and angular velocity and linear acceleration data. In this embodiment, the underground mine environment point cloud data is obtained by laser radar scanning. The laser radar scan obtains one frame of underground mine environment point cloud data at a time. The angular velocity and linear acceleration data of the scraper are collected by an inertial measurement unit. The underground mine environment point cloud data includes multiple frames of underground mine environment point cloud data.
[0019] Preprocessing of underground mine environment point cloud data includes: First, the underground mine environment point cloud data is downsampled to reduce the burden of subsequent calculations. In this embodiment, downsampling is performed by processing the point cloud data using a voxel grid filtering method, which can effectively reduce the density of the point cloud data while preserving important environmental features as much as possible.
[0020] Then, the Euclidean distance clustering algorithm is used to filter the point cloud data to remove sporadic noise caused by dust and ensure the quantity and quality of subsequent processing. Figure 3In Euclidean clustering, distance is determined using Euclidean distance. For a point P in the point cloud data, the k nearest neighbor search algorithm is used to find the k points closest to P. If any of these points have a distance less than a set threshold, they are clustered into set Q. If the number of elements in set Q does not increase, the clustering process ends. Otherwise, points other than P are selected from set Q and the above process is repeated until the number of elements in set Q stops increasing.
[0021] In this embodiment, the SLAM algorithm uses FastLIO, a high-performance laser-inertial odometry (LIO) framework designed specifically for lightweight mobile platforms. Its core advantage lies in breaking through the traditional feature-dependent mapping model and achieving tight coupling and real-time fusion of raw point clouds and inertial data.
[0022] FastLIO uses preprocessed underground mine environment point cloud data and loader pose information as input. Without manually designed feature extraction, it directly uses the original point cloud and a dynamic map constructed using an incremental kd-tree (ikd-Tree) for rigid registration, improving the ability to preserve details in the mine environment. Ultimately, it outputs a global point cloud map of the underground mine environment and real-time odometry information (the odometry information is not needed later).
[0023] While constructing the global point cloud map of the underground mine environment, an initialized SC (Scan Context) database is also constructed.
[0024] Obtain underground mine environment point cloud data and preprocess it. The preprocessing steps are the same as the above preprocessing and will not be described in detail here.
[0025] Generate SC descriptors based on each frame of underground mine environment point cloud data, specifically including: like Figure 5 As shown, a frame of underground mine environment point cloud data is divided into Ns sectors according to the azimuth angle in the laser radar coordinate system, and is divided into Nr rings according to the radius in the laser radar coordinate system. The intersection of each sector and ring is a block area.
[0026] The maximum height of the point cloud in each block is used as the block identification number. For blocks without point clouds, the block identification number is set to 0. The identification numbers of all blocks are combined into a two-dimensional matrix (Nr×Ns), and the two-dimensional matrix is used as the SC descriptor of the underground mine environment point cloud data frame.
[0027] Determine whether each frame of underground mine environment point cloud data is a key frame. If at least one of the following conditions is met, it is considered a key frame. The conditions include: The distance moved by the scraper in the point cloud data of this frame exceeds the preset distance (the preset distance is 1m in this embodiment); The yaw angle of the scraper in the frame of point cloud data changes by more than a set angle (the set angle is 20° in this embodiment); The cosine similarity between the SC descriptor of the frame and the nearest adjacent key frame is lower than a preset threshold (the preset threshold is 0.7 in this embodiment); The change in the proportion of invalid sectors in the frame of point cloud data exceeds a preset threshold (in this embodiment, the preset threshold is 15%).
[0028] When the frame of underground mine environment point cloud data is a key frame, the SC descriptor of the frame of underground mine environment point cloud data is stored in the SC database, and the final SC database is obtained by traversing all the frame point cloud data.
[0029] Next comes the positioning phase. The process of the positioning phase is as follows: Figure 2 As shown, the specific steps include: Real-time underground mine environment point cloud data is obtained and preprocessed. The preprocessing steps are the same as the above preprocessing and will not be described in detail here.
[0030] The preprocessed current frame point cloud data is matched with the global point cloud map of the underground mine environment to obtain the position and matching score of the current frame point cloud data in the global point cloud map of the underground mine environment.
[0031] Determine whether the scraper positioning and matching score drift based on the pose of the current frame point cloud data in the global point cloud map of the underground mine environment. Specifically: First, the matching score is judged. The matching score between the current frame point cloud data and the global point cloud map of the underground mine environment is compared with the preset threshold. If the matching score is less than the preset threshold, it is judged that the scraper positioning drifts. If the matching score is greater than the preset threshold, the position of the current frame point cloud data in the global point cloud map of the underground mine environment is judged: Calculate the pose distance between the pose of the current frame point cloud data in the global point cloud map of the underground mine environment and the pose of the previous frame point cloud data in the global point cloud map of the underground mine environment.
[0032] When the pose distance between the current frame point cloud data in the global point cloud map of the underground mine environment and the pose between the previous frame point cloud data in the global point cloud map of the underground mine environment meets the following constraint conditions, the scraper positioning is judged to have drifted. Otherwise, the scraper positioning does not drift. The constraint condition is expressed as follows: ; in, represents the pose distance, Indicates time interval The average speed within Indicates the interval between obtaining the current frame of point cloud data and obtaining the previous frame of point cloud data.
[0033] When the scraper positioning does not drift, the position of the current frame point cloud data in the global point cloud map of the underground mine environment is directly used as the positioning result.
[0034] When the scraper drifts, the SC descriptor of the current frame point cloud data is generated (the steps are the same as above), and the SC descriptor of the current frame point cloud data is matched with the SC descriptor of the frame point cloud data in the SC database. Specifically: S1. Take the average of the element values of each row (ring) in the matrix of the SC descriptor of each frame point cloud data and the current frame point cloud data in the SC database to obtain a one-dimensional vector, and use the one-dimensional vector as the ring key.
[0035] S2. Traverse the SC descriptor of each frame point cloud data in the SC database and calculate the Euclidean distance between it and the SC descriptor of the current frame point cloud data.
[0036] S3. Select the SC descriptors of the 5 frames of point cloud data with the smallest Euclidean distance.
[0037] S4. Rotate the SC descriptor of each frame point cloud data, and calculate the cosine similarity between the SC descriptor of the rotated frame point cloud data and the SC descriptor of the current frame.
[0038] S5. Compare the cosine similarity of the SC descriptors of the five rotated frame point cloud data with the set threshold. If the cosine similarity is greater than the set threshold, the match is successful. Otherwise, continue S4 to S5 until the rotation angle reaches 360°. If the cosine similarity is still less than the set threshold, the match fails.
[0039] The number of rotations is determined by the number of sectors in the SC descriptor of the point cloud data frame. For example, if Ns is 60, the angle of a sector is 6 degrees, and each rotation is 6 degrees, and the number of rotations is 60 times.
[0040] A successful match indicates a successful relocalization. The rotation angle of the SC descriptor of the frame point cloud data corresponding to the successful match is used as the yaw angle deviation, and the yaw angle deviation is used to correct the current posture.
[0041] If the matching is unsuccessful, it indicates that the positioning phase of the next frame of point cloud data will continue. If the positioning drifts and the matching is unsuccessful for a period of time (1 second in this embodiment), an emergency stop signal for the scraper is issued. Example 2
[0042] Based on Example 1, this embodiment also provides an underground mine environment mapping and positioning system, including a laser radar 1, an inertial measurement unit 2 and a mapping and positioning system. The laser radar 1 is used to collect point cloud data of the underground mine environment, and the inertial measurement unit 2 is used to obtain the position information of the shovel loader.
[0043] In this embodiment, the laser radar 1 is installed above the inertial measurement unit 2 through a bracket, and the inertial measurement unit 2 is installed above the cab of the scraper through a bracket. The laser radar 1 and the inertial measurement unit 2 are both electrically connected to the mapping and positioning system. The mapping and positioning system includes a mapping module, an SC database construction module, a positioning module, and a repositioning module.
[0044] The mapping module is configured to construct a global point cloud map of the underground mine environment.
[0045] The SC database construction module is configured to construct the SC database.
[0046] The positioning module is configured to obtain real-time underground mine environment point cloud data; match the current frame underground mine environment point cloud data with the global point cloud map of the underground mine environment to obtain the pose and matching score of the current frame point cloud data in the global point cloud map of the underground mine environment; judge whether the scraper positioning drifts based on the pose and matching score of the current frame point cloud data in the global point cloud map of the underground mine environment, and when the scraper positioning does not drift, the pose of the current frame point cloud data in the global point cloud map of the underground mine environment is used as the positioning result.
[0047] The repositioning module is configured to generate an SC descriptor of the current frame point cloud data when the scraper loader drifts in positioning, and match the SC descriptor of the current frame point cloud data with the SC descriptor of the frame point cloud data in the SC database. If the match is successful, the rotation angle of the SC descriptor of the frame point cloud data corresponding to the successful match is used as the yaw angle deviation, and the posture of the scraper is corrected and used as the positioning result. If the match fails, it is determined to be positioning drift.
[0048] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for mapping and positioning an underground mine environment, characterized in that: include: Constructing a global point cloud map of the underground mine environment and an SC database, wherein the SC database includes SC descriptors of multiple frames of point cloud data; Obtain real-time underground mine environment point cloud data; Matching the current frame underground mine environment point cloud data with the global point cloud map of the underground mine environment to obtain the pose and matching score of the current frame point cloud data in the global point cloud map of the underground mine environment; Determine whether the scraper positioning is drifting based on the pose and matching score of the current frame point cloud data in the global point cloud map of the underground mine environment: If the positioning of the scraper does not drift, the position of the current frame point cloud data in the global point cloud map of the underground mine environment is used as the positioning result; If the scraper drifts in positioning, an SC descriptor is generated for the current frame point cloud data, and the SC descriptor of the current frame point cloud data is matched with the SC descriptor of the frame point cloud data in the SC database. If the match is successful, the posture is corrected and used as the positioning result. If the match fails, it is determined to be positioning drift.
2. The underground mine environment mapping and positioning method according to claim 1, characterized in that: The method for constructing a global point cloud map of an underground mine environment includes: Obtain underground mine environment point cloud data and scraper angular velocity and linear acceleration data; The underground mine environment point cloud data and the angular velocity and linear acceleration data of the scraper are used as input, and the global point cloud map of the underground mine environment is obtained based on the SLAM algorithm output.
3. The underground mine environment mapping and positioning method according to claim 1, characterized in that: The method for constructing the SC database includes: Build SC database; Acquire point cloud data of underground mine environment; Generate SC descriptors based on each frame of underground mine environment point cloud data; It is determined whether each frame of underground mine environment point cloud data is a key frame. If it is a key frame, the SC descriptor of the frame of underground mine environment point cloud data is stored in the SC database.
4. The underground mine environment mapping and positioning method according to claim 3, characterized in that: Generating the SC descriptor according to each frame of underground mine environment point cloud data includes: A frame of underground mine environment point cloud data is divided into a plurality of sectors according to azimuth angles, and is divided into a plurality of rings according to radius, and the intersection of each sector and ring is used as a block area; The maximum height of the point cloud in each block is used as the identification number of the block. For blocks without point clouds, the identification number of the block is set to 0. The identification numbers of all blocks are combined into a two-dimensional matrix, and the two-dimensional matrix is used as the SC descriptor of the underground mine environment point cloud data of this frame; The azimuth angle and radius are determined by the laser radar coordinate system.
5. The underground mine environment mapping and positioning method according to claim 3, characterized in that: The underground mine environment point cloud data is judged as a key frame only if it meets at least one of the following conditions: The moving distance of the scraper in this frame of point cloud data exceeds the preset distance; The yaw angle of the scraper in this frame of point cloud data exceeds the set angle; The cosine similarity between the SC descriptor of the frame and the nearest adjacent key frame is lower than a preset threshold; The change in the proportion of invalid sectors in the point cloud data of this frame exceeds a preset threshold.
6. The underground mine environment mapping and positioning method according to any one of claims 1 to 3, characterized in that: After obtaining the real-time underground mine environment point cloud data and the underground mine environment point cloud data, pre-processing is also performed on the data. The pre-processing includes: The underground mine point cloud data is processed using voxel grid filtering to reduce the point cloud density and obtain the downsampled underground mine point cloud data. The downsampled underground mine point cloud data is filtered using the Euclidean distance clustering method to obtain preprocessed real-time underground mine environment point cloud data and underground mine environment point cloud data.
7. The underground mine environment mapping and positioning method according to claim 1, characterized in that: The determining whether the positioning of the scraper is drifting according to the pose and matching score of the current frame point cloud data in the global point cloud map of the underground mine environment includes: Compare the matching score between the current frame point cloud data and the global point cloud map of the underground mine environment with a preset threshold. If the matching score is less than the preset threshold, it is determined that the scraper positioning is drifting. If the matching score is greater than a preset threshold, the pose distance between the pose of the current frame point cloud data in the global point cloud map of the underground mine environment and the pose of the previous frame point cloud data in the global point cloud map of the underground mine environment is calculated; When the pose distance between the pose of the current frame point cloud data in the global point cloud map of the underground mine environment and the pose of the previous frame point cloud data in the global point cloud map of the underground mine environment meets the following constraint conditions, it is determined that the scraper is drifting in positioning; otherwise, it is determined that the scraper is not drifting in positioning. The expression of the constraint conditions is: ; in, represents the pose distance, Indicates time interval The average speed within Indicates the interval between obtaining the current frame of point cloud data and obtaining the previous frame of point cloud data.
8. The underground mine environment mapping and positioning method according to claim 1, characterized in that: The matching of the SC descriptor of the current frame point cloud data with the SC database includes: S1. averaging the values of each row of the matrix of the SC descriptors of each frame of point cloud data in the SC database and the current frame of point cloud data to obtain a one-dimensional vector, and using the one-dimensional vector as the ring key; S2. Traverse the ring key of each frame point cloud data in the SC database and calculate the Euclidean distance between it and the ring key of the current frame point cloud data; S3, select the SC descriptor of the 5 frames of point cloud data with the smallest Euclidean distance; S4, rotating the SC descriptor of each frame point cloud data, and calculating the cosine similarity between the SC descriptor of the rotated frame point cloud data and the SC descriptor of the current frame; S5. Compare the cosine similarity of the SC descriptors of the five rotated frame point cloud data with the set threshold. If any cosine similarity is greater than the set threshold, the match is successful. Otherwise, continue S4 to S5 until the rotation angle reaches 360°. If the cosine similarity is still less than the set threshold, the match fails. The rotation angle corresponding to the successful matching is used as the yaw angle deviation, and the yaw angle deviation is used for posture correction.
9. The underground mine environment mapping and positioning method according to claim 8, characterized in that: The number of rotations is determined by the number of sectors of the SC descriptor of the frame point cloud data, and each rotation angle is obtained by evenly dividing 360° according to the number of sectors of the SC descriptor of the frame point cloud data.
10. An underground mine environment mapping and positioning system, characterized in that: Used to implement the underground mine environment mapping and positioning method according to any one of claims 1 to 9, which comprises a laser radar, an inertial measurement unit and a mapping and positioning system, wherein the laser radar is used to collect point cloud data of the underground mine environment, and the inertial measurement unit is used to obtain angular velocity and linear acceleration data of the scraper; The laser radar is arranged above the inertial measurement unit, and the inertial measurement unit is arranged above the cab of the scraper through a bracket. The laser radar and the inertial measurement unit are both electrically connected to the mapping and positioning system. The mapping and positioning system includes a mapping module, an SC database construction module, a positioning module and a repositioning module. The mapping module is configured to construct a global point cloud map of the underground mine environment; The SC database construction module is configured to construct the SC database; a positioning module configured to obtain real-time underground mine environment point cloud data; match the current frame underground mine environment point cloud data with the global point cloud map of the underground mine environment to obtain the position and posture of the current frame point cloud data in the global point cloud map of the underground mine environment; determine whether the positioning of the scraper loader has drifted based on the position and posture of the current frame point cloud data in the global point cloud map of the underground mine environment and the matching score; and if the positioning of the scraper loader has not drifted, use the position and posture of the current frame point cloud data in the global point cloud map of the underground mine environment as the positioning result; The repositioning module is configured to generate the SC descriptor of the current frame point cloud data when the scraper drifts in positioning, and match the SC descriptor of the current frame point cloud data with the SC descriptor of the frame point cloud data in the SC database. If the match is successful, the posture is corrected and used as the positioning result. If the match fails, it is determined to be positioning drift.
