A mine area obstacle detection method, system, electronic device and readable storage medium

Abstract

The application discloses a kind of mine area obstacle detection method, system, electronic equipment and readable storage medium, comprising: obtaining time-synchronized 4D light field camera point cloud data and laser radar point cloud data, according to the different position information in two sensors, solve out calibration parameter;Find each the near neighbor point of the laser radar point cloud data in 4D light field point cloud data, and whether it is abnormal point according to adjacent point number judgment;Then respectively obtain 4D light field camera and laser radar's obstacle point cloud, are re-registered to obtain new calibration parameter;Corresponding is carried out to update boundary box information, and the number of abnormal points is higher than the threshold value Obstacle point cloud filtering;Using 4D light field camera and laser radar detection characteristics reach high accuracy mine area target detection, using 4D light field camera can penetrate the characteristics of dust barrier, to the obstacle finally detected, to improve the range and precision of mine area obstacle detection Abnormal point number judgment.

Description

Technical Field

[0001] This invention relates to the field of sensor target detection technology, and more specifically, to a method, system, electronic device, and readable storage medium for detecting obstacles in mining areas. Background Technology

[0002] Obstacle detection technology has important applications in the field of autonomous driving, helping vehicles and drivers to detect obstacles ahead and take avoidance actions in advance. Current obstacle detection methods mainly include camera-based visual obstacle detection, LiDAR-based obstacle detection, and obstacle detection methods based on other sensors. Mining environments are complex, with numerous irregular obstacles. Furthermore, vehicles in mining areas are large, and single sensors are prone to detection blind spots, necessitating the fusion of multiple sensors for comprehensive detection.

[0003] LiDAR point cloud data is characterized by its disorder, irregularity, and low resolution, making it difficult to process and analyze. However, it can provide precise 3D information about the scene. Imaging devices, on the other hand, lack 3D information in their images. Therefore, using images alone for 3D target detection is less effective than using LiDAR point cloud data alone. However, LiDAR point cloud data possesses RGB information, which LiDAR point cloud data lacks, and the two types of data can complement each other in terms of features. Therefore, researching 3D target detection for autonomous vehicles based on the fusion of image and LiDAR point cloud data can improve detection accuracy.

[0004] Chinese patent CN113963335A, entitled "Road Obstacle Detection Method Based on Image and Point Cloud Data," provides a road obstacle detection method based on image and point cloud data. This invention compensates for the lack of three-dimensional spatial coordinate information in two-dimensional images by reconstructing the three-dimensional structure of objects. Then, it learns the feature representations containing three-dimensional structural information to predict whether the object is an obstacle. However, this method requires observation of obstacles from multiple angles and is not effective for obstacle detection in autonomous vehicles.

[0005] Chinese patent CN114782922A, entitled "Obstacle Detection Method, Apparatus, and Electronic Equipment," provides an obstacle detection method, apparatus, and electronic equipment. Relating to the field of artificial intelligence, the method includes: acquiring a 3D point cloud and a 2D image corresponding to the environment in which a target vehicle is located; randomly selecting multiple pixels from the 2D image; determining a target pixel from the multiple pixels based on the 3D point cloud, wherein the target pixel is the pixel corresponding to an obstacle in the environment in which the target vehicle is located; performing back-projection processing on the target pixel to obtain a target 3D point cloud, wherein the target 3D point cloud is used to supplement the number of points in the 3D point cloud; and detecting the 3D point cloud and the target 3D point cloud to obtain obstacle information. The random pixel selection method of this method is difficult to determine the target pixel in similar scenes in mining areas, and may not achieve ideal results in obstacle detection in mining areas.

[0006] For obstacle detection methods that fuse camera and LiDAR data, the common approach is to map point cloud data onto an image for processing. The accuracy of this mapping significantly impacts the fusion detection results. Mining environments are complex and rugged, and the movement of autonomous vehicles in these environments can easily cause sensor placement shifts, altering the mapping. Furthermore, the harsh environment of mining areas makes it easy to misidentify dust and other debris as obstacles. Summary of the Invention

[0007] This invention aims to provide a method and system for obstacle detection in mining areas, addressing the challenges posed by the complex environment and rugged terrain of mining environments. These challenges include the tendency for sensor installation positions to shift when unmanned vehicles operate in these environments, and the reduction in sensor detection capabilities due to the large amounts of dust generated. The goal is to improve the range and accuracy of obstacle detection in mining areas.

[0008] In view of this, the first aspect of the present invention is to provide a method for detecting obstacles in a mining area.

[0009] A second aspect of the present invention is to provide an obstacle detection system for mining areas.

[0010] A third aspect of the present invention is to provide an electronic device.

[0011] A fourth aspect of the present invention is to provide a computer-readable storage medium.

[0012] The first aspect of the present invention provides a method for obstacle detection in a mining area, comprising the following steps: S1, acquiring time-synchronized 4D light field camera point cloud data and lidar point cloud data during vehicle operation; S2, solving for calibration parameters of the 4D light field camera and lidar based on the different position information of the calibration object in the two sensor coordinate systems; S3, projecting the lidar point cloud data onto the 4D light field camera coordinate system according to the calibration parameters, finding the nearest neighbor points of each lidar point cloud data in the 4D light field point cloud data, and determining whether the current lidar point cloud data is an anomaly based on the number of nearest neighbor points; 4. Based on the 4D light field camera point cloud data, perform target detection using a convolutional neural network to obtain the obstacle point cloud of the 4D light field camera; S5. Based on the lidar point cloud data, perform ground point cloud filtering and non-ground point cloud clustering to obtain the obstacle point cloud of the lidar; S6. Re-register the obstacle point clouds obtained by the lidar and the 4D light field camera to obtain new calibration parameters; S7. Based on the new calibration parameters, match the obstacle point clouds of the lidar and the 4D light field camera to update the bounding box information, and filter out obstacle point clouds with anomalies exceeding a threshold to complete obstacle detection.

[0013] Specifically, step S4 includes: S401, inputting the image from the 4D light field camera into a convolutional neural network to obtain the coordinates of the detection box position in the image; S402, mapping the coordinates of the detection box position to the coordinates of the 4D light field camera point cloud, and clustering the data of the 4D light field camera point cloud to obtain the obstacle point cloud of the 4D light field camera.

[0014] In addition, the technical solutions provided by embodiments of the present invention may also have the following additional technical features:

[0015] In any of the above technical solutions, when the 4D light field camera and the lidar are installed on the vehicle, the detection range of the lidar covers the detection blind zone of the 4D light field camera.

[0016] In any of the above technical solutions, step S3 specifically includes: S301, transforming the LiDAR point cloud data according to the calibration parameters obtained in S2 to obtain LiDAR point cloud data in the 4D light field camera coordinate system; S302, traversing each LiDAR point cloud data and performing a nearest neighbor search in the 4D light field camera point cloud using a Kdtree to search for nearest neighbor points within a preset distance range; S303, determining the size of the number of nearest neighbor points of the current LiDAR point cloud data compared to a threshold; if the value is large, the current LiDAR point cloud data is considered an anomaly; wherein, the preset distance is 0.05m.

[0017] In any of the above technical solutions, step S5 specifically includes: S501, performing planar fitting on the lidar point cloud and filtering out the largest in-plane point cloud to obtain a non-ground point cloud; S502, clustering the non-ground point cloud to obtain the lidar obstacle point cloud.

[0018] In any of the above technical solutions, the plane fitting adopts the RANSAC method, and the clustering steps include: randomly extracting a point as the center point, searching for the nearest neighbor points of the center point around it; after finding a nearest neighbor point, using the current nearest neighbor point as the search center point to perform iterative search until no new nearest neighbor points can be found.

[0019] In any of the above technical solutions, step S6 specifically includes: S601, re-registering based on the calibration parameters according to the geometric coordinates of the obstacle point cloud obtained in S4 and S5; S602, prioritizing pose transformation of non-abnormal points to obtain the new calibration parameters and updating the calibration parameters obtained in S2.

[0020] In any of the above technical solutions, step S7 specifically includes: S701, obtaining the correspondence of the obstacle point clouds based on the obstacle point clouds obtained in S4 and S5 using the new calibration parameters; S702, comparing the geometric values ​​of the corresponding obstacle point clouds, selecting the larger value among the geometric values ​​as new geometric information, and obtaining a complete obstacle point cloud; S703, counting the number of corresponding obstacle point clouds and the number of each with anomalies, filtering out obstacle point clouds with anomalies exceeding a threshold, and completing obstacle detection; wherein, the geometric values ​​include length, width, and height.

[0021] The first aspect of the present invention provides a method for obstacle detection in mining areas, comprising: a data acquisition module for acquiring 4D light field camera data and lidar point cloud data from an autonomous vehicle; a calibration module for spatially calibrating the 4D light field camera data and lidar point cloud data, and recalibrating the spatial position relationship of the sensors based on the detection results after obstacle detection; an anomaly detection module for projecting the lidar point cloud data onto the 4D light field camera data and detecting anomalies; a camera detection module for performing target detection on the 4D light field camera data using a deep learning method; a radar detection module for performing target detection based on the lidar point cloud data; and a fusion module for fusing the obstacle position and bounding box information obtained by the camera detection module and the radar detection module, and filtering out abnormal obstacles; wherein the obstacle detection system for mining areas is used to implement the obstacle detection method for mining areas described in the first aspect of the technical solution.

[0022] A third aspect of the present invention provides an electronic device comprising: a memory, a processor, and a program stored in the memory and executable on the processor, the processor executing the program to implement the above-described obstacle detection method based on the fusion of lidar and 4D light field camera.

[0023] This invention provides an electronic device equipped with a positioning system for feature degradation scenarios as described in the above-mentioned technical solution. Therefore, the electronic device proposed in this invention possesses all the beneficial effects of the positioning system for feature degradation scenarios described in the above-mentioned technical solution, which will not be elaborated further here.

[0024] A fourth aspect of the present invention provides a computer-readable storage medium storing a program that is executed by a processor to implement an obstacle detection method based on the fusion of lidar and 4D light field camera.

[0025] This invention provides a computer-readable storage medium, which includes any medium capable of storing or transmitting information. Examples of readable storage media include electronic circuits, semiconductor memory electronics, read-only memory (ROND), random access memory (RAMD), compact disc read-only memory (CD-ROMD), flash memory, erasable ROM (EROMD), magnetic tape, floppy disk, optical disk, hard disk, fiber optic media, radio frequency (RF) links, optical data storage electronics, etc. The code segments can be downloaded via computer networks such as the Internet and intranets.

[0026] The beneficial effects of this invention compared to the prior art are as follows:

[0027] This invention proposes an obstacle detection system for mining areas, which integrates a 4D light field camera and a lidar, utilizing the detection characteristics of both to achieve highly accurate target detection in mining areas;

[0028] This invention proposes an anomaly detection method for obstacles in mining areas. It utilizes the characteristic of a 4D light field camera that can see through dust obstacles, projects the lidar point cloud into the camera coordinate system, and compares it point by point with the three-dimensional coordinate information in the 4D light field camera. Anomalies are marked, and the number of anomalies in the finally detected obstacles is judged. When the number of anomalies reaches a certain threshold, it is marked as an abnormal obstacle.

[0029] This invention proposes an online calibration method for 4D light field cameras and lidar. It extracts feature points from the point cloud information in the obstacle detection results of the 4D light field camera to improve the calibration speed. Based on the prior calibration pose, it registers with the lidar point cloud data and updates the calibration parameters.

[0030] Additional aspects and advantages of embodiments of the invention will become apparent in the following description or may be learned by practice of embodiments of the invention. Attached Figure Description

[0031] The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of the invention.

[0032] Figure 1 This is a flowchart of the obstacle detection process based on the fusion of lidar and 4D light field camera of the present invention.

[0033] Figure 2 This is a diagram of the obstacle detection module based on the fusion of lidar and 4D light field camera of the present invention. Detailed Implementation

[0034] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0035] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0036] An embodiment of the first aspect of the present invention provides a method for detecting obstacles in mining areas. In some embodiments of the present invention, such as... Figure 1 As shown, a method for detecting obstacles in mining areas is provided. This method for detecting obstacles in mining areas includes the following steps S1 to S7:

[0037] Step 1: Install LiDAR sensors and 4D light field cameras on unmanned vehicles in the mining area. During installation, ensure that the detection range of the LiDAR covers the blind spot of the 4D light field camera, and allow the vehicle to collect data from multiple sensors.

[0038] Step 2: Spatiotemporally synchronize the data from each sensor: Set up a calibration object in an open scene, and calculate the installation parameters of the two sensors by using the position of the calibration object in the LiDAR sensor and the 4D light field camera.

[0039] Specifically, the position parameters of the calibration object in the LiDAR point cloud are set to (x, y, z), and the position parameters of the calibration object in the 4D light field camera are set to (x, y, z). ’ ,y ’ ,z ’ Then the following formula applies:

[0040]

[0041] The rotation matrix R and translation matrix T are calculated using multiple corresponding calibration parameters.

[0042] Step 3: Based on the rotation matrix R and translation matrix T calculated in Step 2, transform the lidar point cloud into the 4D light field camera point cloud coordinate system;

[0043] Traverse each point in the lidar point cloud and use Kdtree to quickly search for nearest neighbor points within a certain distance range in the 4D light field camera point cloud. If the number of nearest neighbor points does not meet the threshold requirement, mark the point cloud as an abnormal point cloud.

[0044] Specifically, the value for a certain distance is 0.05m, and the threshold for the number of nearest neighbors is determined using the following formula:

[0045]

[0046] Where n is the threshold for the number of nearest neighbors, l is the search distance, d is the distance of the point from the lidar, and dx,dy represent the horizontal and vertical angular resolutions of the lidar.

[0047] Step 4: Input the image information from the 4D light field camera into the convolutional neural network using a pre-trained neural network model to obtain the location of the detection box in the image;

[0048] The pixel coordinates are mapped to the point cloud data coordinates, and the 4D light field point cloud data is clustered again to obtain the coordinate information for obstacle detection by the camera.

[0049] Step 5: Perform plane fitting on the LiDAR point cloud. Use the RANSAC method to fit the plane, filter out the largest point cloud in the plane, and perform Euclidean clustering on the remaining non-ground point cloud to obtain the coordinate system information for obstacle detection by the LiDAR.

[0050] Specifically, a point is randomly selected from the non-terrestrial point cloud as the center point. Within a given neighborhood, its nearest neighbors are searched. Once a nearest neighbor is found, the search continues iteratively, using this new nearest neighbor as the origin. When no new points are found, clustering is complete.

[0051] Step 6: Re-register the obstacle point clouds acquired by the LiDAR and 4D light field camera using the ICP matching method, based on prior calibration parameters. The optimization criterion for matching is minimizing the Euclidean clustering between corresponding points. After solving for the new calibration parameters, update the parameters.

[0052] Step 7: Based on the new calibration parameters obtained in Step 6, match the obstacles in the LiDAR point cloud with the obstacles in the 4D light field camera point cloud. Compare the geometric positional relationships of the corresponding obstacles and update the bounding box information. Simultaneously, count the outliers in each obstacle's point cloud. When the number of outliers exceeds a threshold, the obstacle is considered an abnormal obstacle, possibly a false detection by the LiDAR due to dust or other factors. Filter out this obstacle, completing the final obstacle target detection.

[0053] Further, step S2 specifically involves: selecting a clearly marked calibration object in an open scene, such as placing a calibration object with high reflectivity; comparing and calculating the location information of the calibration object in the point cloud data of the 4D light field camera data and the location information of the calibration object in the point cloud data of the lidar to obtain accurate calibration data for the 4D light field camera and lidar.

[0054] Further, step S3 specifically involves: using the calibration parameters obtained in S2, transforming the coordinate system of the points in the point cloud data to obtain the lidar point cloud data in the 4D light field camera coordinate system. Each lidar point cloud is traversed, and a Kdtree is used to perform a nearest neighbor search within the 4D light field camera point cloud, searching for nearest neighbors within a certain distance. If the number of 4D light field camera point clouds near a given lidar point cloud does not meet a certain threshold, the lidar point cloud is considered potentially an anomaly and is marked.

[0055] Further, step S5 specifically involves mapping the obstacle location information obtained in S4 onto the lidar coordinate system to obtain the approximate location of the obstacles in the lidar point cloud, and then performing clustering.

[0056] Further, step S6 specifically involves: based on the final geometric coordinates of the obstacle point cloud obtained in S5, re-registering is performed based on the prior pose transformation relationship. During registration, the pose transformation of non-anomaly points is prioritized. After obtaining the new pose transformation relationship, the obstacle calibration parameters obtained in S2 are updated.

[0057] Further, step S7 specifically involves the following process: Based on the obstacle information from the two point clouds obtained in S6, since both have been calibrated to the same coordinate system, the correspondence is obtained by comparing the coordinate positions of the obstacles in the two points. After finding the corresponding obstacle, the geometric values ​​of the two points are compared, such as length, width, and height. The larger value is selected as the new geometric information, ultimately obtaining complete obstacle information. Simultaneously, the number of point clouds for each corresponding obstacle and the number of abnormal point clouds are counted. If the number of point clouds is too small or the proportion of abnormal point clouds to the total number is too large, the obstacle is considered an abnormal obstacle and is filtered out.

[0058] A second aspect of the present invention provides an obstacle detection system for mining areas. In some embodiments of the present invention, such as... Figure 2As shown, an obstacle detection system for mining areas is proposed. This system includes: an acquisition module, a calibration module, a preprocessing module, a segmentation module, a clustering module, and a fusion module. Specifically: the acquisition module acquires 4D light field camera data and LiDAR point cloud data from autonomous vehicles; the calibration module performs spatial calibration on the 4D light field camera data and LiDAR point cloud data, and performs real-time calibration during detection; the anomaly detection module projects the LiDAR point cloud data onto the 4D light field camera data and detects anomalies; the camera detection module performs target detection on the 4D light field camera data using deep learning methods; and the LiDAR detection module performs target detection based on the LiDAR point cloud data. The fusion module fuses the obstacle position and bounding box information obtained from the camera detection module and the radar detection module, and filters out abnormal obstacles. This obstacle detection system for mining areas is used to implement the obstacle detection method for mining areas in any of the above embodiments.

[0059] A third aspect of the present invention provides an electronic device. In some embodiments of the present invention, an electronic device is proposed, comprising: a memory storing a program or instructions; and a processor executing the program or instructions to implement a mine obstacle detection method according to any embodiment. Therefore, the electronic device proposed in this embodiment possesses all the beneficial effects of the mine obstacle detection method defined in the above embodiments, and will not be elaborated further here.

[0060] A fourth aspect of the present invention provides a readable storage medium. In some embodiments of the present invention, a readable storage medium is proposed that stores a program or instructions, which, when executed by a processor, implement the steps of the ship information display method in any of the above embodiments. Therefore, the readable storage medium proposed in this embodiment possesses all the beneficial effects of the ship information display method in any of the above embodiments, and will not be elaborated further here.

[0061] In specific embodiments, a readable storage medium can include any medium capable of storing or transmitting information. Examples of readable storage media include electronic circuits, semiconductor memory devices, read-only memory (ROM), random access memory (RAM), compact disc read-only memory (CD-ROM), flash memory, erasable ROM (EROM), magnetic tape, floppy disk, optical disk, hard disk, fiber optic media, radio frequency (RF) links, optical data storage devices, etc. Code segments can be downloaded via computer networks such as the Internet or intranets.

[0062] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this invention, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0063] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for detecting obstacles in a mining area, characterized in that, Includes the following steps: S1, during vehicle operation, acquires time-synchronized 4D light field camera point cloud data and lidar point cloud data; S2, Based on the different position information of the calibration object in the coordinate system of the 4D light field camera and the lidar, solve for the calibration parameters of the 4D light field camera and the lidar; S3. Based on the calibration parameters, project the lidar point cloud data into the coordinate system of the 4D light field camera, find the nearest neighbor points of each lidar point cloud data in the 4D light field point cloud data, and determine whether the current lidar point cloud data is an abnormal point based on the number of nearest neighbor points. S4, Based on the point cloud data of the 4D light field camera, use a convolutional neural network to perform target detection and obtain the obstacle point cloud of the 4D light field camera; S5, Based on the lidar point cloud data, perform ground point cloud filtering and non-ground point cloud clustering to obtain the lidar obstacle point cloud; S6 re-registers the obstacle point cloud acquired by the lidar and 4D light field camera to obtain new calibration parameters; S7. Based on the new calibration parameters, the obstacle point clouds of the LiDAR and the 4D light field camera are matched to update the bounding box information, and the obstacle point clouds with anomalies exceeding the threshold are filtered out to complete obstacle detection.

2. The method for detecting obstacles in a mining area according to claim 1, characterized in that, When the 4D light field camera and lidar are installed on the vehicle, the detection range of the lidar covers the detection blind zone of the 4D light field camera.

3. The method for detecting obstacles in a mining area according to claim 1, characterized in that, Step S3 specifically includes: S301, based on the calibration parameters obtained in S2, the lidar point cloud data is transformed to obtain lidar point cloud data in the 4D light field camera coordinate system. S302, traverse each of the lidar point cloud data, and use Kdtree to perform a nearest neighbor search in the 4D light field camera point cloud to search for nearest neighbor points within a preset distance range; S303, determine the number of nearest neighbor points in the current lidar point cloud data and the value of the threshold. If the value is large, the current lidar point cloud data is an abnormal point. The preset distance is 0.05m.

4. The method for detecting obstacles in a mining area according to claim 1, characterized in that, Step S5 specifically includes: S501 performs planar fitting on the lidar point cloud and filters out the largest point cloud in the plane to obtain the non-ground point cloud. S502, cluster the non-ground point cloud to obtain the obstacle point cloud of the lidar.

5. The method for detecting obstacles in a mining area according to claim 4, characterized in that, The plane fitting employs the RANSAC method, and the clustering steps include: Randomly select a point as the center point, and search for the nearest neighbor points of the center point around it; Once a nearest neighbor is found, the search continues iteratively, using the current nearest neighbor as the search center, until no new nearest neighbor can be found.

6. The method for detecting obstacles in a mining area according to claim 1, characterized in that, Step S6 specifically includes: S601, Based on the geometric coordinates of the obstacle point cloud obtained in S4 and S5, re-register based on the calibration parameters; S602, prioritize the pose transformation of non-abnormal points to obtain the new calibration parameters, and update the calibration parameters obtained in S2.

7. The method for detecting obstacles in a mining area according to claim 1, characterized in that, Step S7 specifically includes: S701, Based on the obstacle point cloud obtained in S4 and S5, the correspondence of the obstacle point cloud is obtained using the new calibration parameters; S702, compare the geometric values ​​of the corresponding obstacle point clouds, and select the larger value as the new bounding box information to obtain the complete obstacle point cloud; S703, count the number of corresponding obstacle point clouds and the number of each with abnormal points, filter out obstacle point clouds with an abnormal number of points higher than the threshold, and complete obstacle detection; The geometric values ​​include length, width, and height.

8. An obstacle detection system for mining areas, characterized in that, include: The acquisition module is used to acquire 4D light field camera data and lidar point cloud data of autonomous vehicles; The calibration module is used to spatially calibrate the 4D light field camera data and lidar point cloud data, and after obstacle detection, to recalibrate the spatial position relationship of the sensors based on the detection results. An anomaly detection module is used to project the lidar point cloud data onto the 4D light field camera data and perform anomaly detection. The camera detection module is used to perform target detection on 4D light field camera data using deep learning methods to obtain obstacle positions and bounding box information in the 4D light field camera. The radar detection module is used to detect targets based on lidar point cloud data and obtain the location and bounding box information of obstacles in the point cloud. The fusion module is used to fuse the obstacle position and bounding box information obtained by the camera detection module and the radar detection module, and to filter out abnormal obstacles. The obstacle detection system for mining areas is used to implement the obstacle detection method for mining areas as described in any one of claims 1-7.

9. An electronic device, characterized in that, include: Memory, which stores programs or instructions; A processor that executes the program or instructions to implement a mining obstacle detection method as described in any one of claims 1-7.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions which are executed by a processor to implement a mining obstacle detection method as described in any one of claims 1-7.

Images