Transparent obstacle detection and map reconstruction method and system based on laser radar point cloud data

Abstract

The invention discloses a transparent obstacle detection and map reconstruction method and system based on laser radar point cloud data, and the method comprises the steps: scanning a target region through a laser radar, and obtaining a plurality of groups of point cloud data; registering the obtained multiple groups of point cloud data to the same coordinate system to form a point cloud data set; dividing a point cloud data set in a dynamic window sampling mode, performing geometric structure analysis, and searching point cloud data conforming to transparent obstacle features; constructing and training an identification model, identifying the point cloud data conforming to the transparent obstacle features by using the identification model, and performing data modification according to an identification result; and replacing the corresponding original data in the original point cloud data set with the modified output data, and constructing a grid map of the target area according to the original data. The problem that in the existing mapping process, due to refraction, diffuse reflection and the like of indoor glass doors and windows or wall surfaces on laser, obstacles cannot be correctly recognized, and the driving safety of unmanned vehicles is affected can be solved.

Description

Technical Field

[0001] The present invention relates to the field of unmanned driving technology, and in particular to a transparent obstacle detection and map reconstruction method and system based on laser radar point cloud data. Background Art

[0002] With the development of artificial intelligence technology and unmanned vehicle sensor technology, unmanned vehicles are increasingly used in various fields of social life. When unmanned vehicles are facing unknown environments, simultaneous localization and mapping (SLAM) has become the core technology for realizing autonomous movement of unmanned vehicles. At present, SLAM technology is divided into two categories according to the sensors it relies on: laser SLAM based on lidar and visual SLAM based on single / binocular cameras. In indoor application scenarios, unmanned vehicles generally use laser SLAM as the main method, because the ranging of lidar is simple and stable, and unmanned vehicles can plan and navigate paths based on the point cloud information returned by the radar to meet the basic needs of mobility.

[0003] However, LiDAR faces challenges in mapping indoor environments, especially for the detection of transparent obstacles such as glass windows or walls. Since LiDAR relies on the reflection of laser signals, the optical properties of transparent obstacles will cause the transmission of laser signals, resulting in most of them being unable to return to the LiDAR receiving sensor, thus affecting the accuracy of indoor environment map construction and making it difficult for vehicles to move. Currently, for unmanned vehicles in indoor environments, in order to meet their low power consumption requirements, they are usually only equipped with single-line LiDAR for mapping and navigation, which means that point cloud information is even scarcer, which further poses challenges to the analysis of point cloud data.

[0004] The advantage is that even if most of the laser signal penetrates the glass object, a small part of the signal can still return to the receiver, thus providing certain conditions for distinguishing the glass object. Therefore, how to extract the characteristics of the glass object from the small amount of return signals and mark it as an obstacle when building the map, so that the unmanned vehicle based on the single-line laser radar can work normally in the indoor environment, has become a difficult problem that needs to be solved urgently. Summary of the invention

[0005] The present invention aims to solve one of the technical problems existing in the related art at least to a certain extent.

[0006] One object of the present invention is to provide a transparent obstacle detection and map reconstruction method based on lidar point cloud data to solve the problem that lidar is not effective in transparent obstacle detection and ensure that unmanned vehicles can operate normally in an environment with transparent obstacles such as doors, windows or walls.

[0007] Another object of the present invention is to provide a transparent obstacle detection and map reconstruction system based on lidar point cloud data.

[0008] In order to achieve the above-mentioned purpose, the present invention provides a transparent obstacle detection and map reconstruction method based on laser radar point cloud data, comprising: S1. The unmanned vehicle uses laser radar to scan the entire target area including transparent obstacles to obtain multiple sets of point cloud data; S2, using a registration algorithm to register the acquired multiple sets of point cloud data to the same coordinate system to form a point cloud data set; S3, dividing the point cloud data set by dynamic window sampling, and performing geometric structure analysis to find point cloud data that meets the characteristics of transparent obstacles; S4. Construct and train a recognition model for identifying transparent obstacles, and use the recognition model to identify the point cloud data that meets the characteristics of transparent obstacles obtained in step S3. If a part is identified as having a transparent obstacle, the point cloud density is supplemented in the output data to be equivalent to that of an ordinary obstacle, and the part is regarded as an ordinary obstacle; if a part is identified as not having a transparent obstacle, the original sparse point cloud is removed from the output data, and the part is regarded as having no obstacle; S5. Replace the corresponding original data in the original point cloud data set with the modified output data of step S4, and construct a grid map of the target area based on the modified output data.

[0009] A further preferred technical solution of the present invention is that the point cloud data set is divided by the local dynamic window sampling method in step S3, and a geometric structure analysis is performed to find point cloud data that meets the characteristics of transparent obstacles. The specific method is: S31, setting the size of the initial window, and traversing the point cloud data set with the initial window; S32, performing geometric structure analysis on the point cloud data in the current window, and finding out the point clouds with geometric features of transparent obstacles. The geometric features of transparent obstacles are as follows: the point clouds at the front and back ends of the data are relatively dense, which are reflected as the edges of transparent obstacles; there are fewer and discontinuous point clouds in the middle, which are reflected as transparent obstacles; if the number of point clouds in the straight line area fitted by the coordinates of the two edges is less than a given threshold, it is considered that the part of the point cloud data is suspected to be a transparent obstacle; S33, increase the window size, and perform the next traversal of the point cloud data set until all suspected transparent obstacles of different sizes are completely detected.

[0010] Preferably, in step S3, each time the window completes sampling of the point cloud data set, the density gradient between adjacent sampling windows is calculated. If the window is sampled for the first time or the gradient density changes exceed the feature loss threshold, the magnification factor To expand the window, the magnification factor The calculation formula is:

[0011] in, is an adjustable coefficient used to adjust the weight of the gradient density in the formula; The calculated magnification factor , set the new window size , The calculation formula is:

[0012] In the formula, The original window size.

[0013] Preferably, the density gradient The calculation method is as follows: the number of point clouds within each sampling window is counted as the local density of the window, and the density gradient is obtained by calculating the density difference between adjacent windows.

[0014] Preferably, the recognition model for identifying transparent obstacles constructed in step S4 includes: The input layer obtains the data input from the sampling window and processes it into a tensor format before passing it to the feature extraction layer; The feature extraction layer first uses a multi-layer perceptron to independently extract features from each point cloud tensor, then obtains a global feature vector of fixed length through maximum pooling and fuses it with local features. The output dimension is The characteristic matrix of Represents the number of point clouds input into the network for each window, Indicates the dimension of the coordinate system; The output layer takes the feature matrix as input and outputs the probability of containing transparent obstacles in the point cloud after processing by the fully connected layer and the softmax function. If the probability exceeds the threshold, it is determined that there is a point cloud with transparent glass obstacles.

[0015] Preferably, the method for training the recognition model is: annotating the point cloud data set detected by the radar to generate a training set, training the recognition model multiple times with the training set, adjusting the model parameters to the optimal values, and the loss function of the training process is the cross entropy loss and the Dice loss.

[0016] Preferably, the data inputted by the sampling window in the input layer is preprocessed into a specified data length, and the preprocessing method is one of the farthest point sampling algorithm and the average sampling algorithm; When the output layer outputs data, it uses the reverse processing method of preprocessing to restore the length of the output data to the normal data length.

[0017] Preferably, the specific method of step S1 is: S11. Planning a path for the unmanned vehicle so that the driving route of the unmanned vehicle can cover the entire target area; S12. The unmanned vehicle uses a laser radar to scan at a set scanning point, and obtains a set of point cloud data after each 360° scan; the point cloud data at least includes distance information and point cloud data intensity information, and the scanning resolution is determined by the radar model.

[0018] Preferably, after constructing the grid map, if there is a dynamic change in the opening or closing state of the glass door, return to step S2, correct the point cloud data acquired in real time to the point cloud data set, re-identify the transparent obstacles, and update the target area grid map.

[0019] Another aspect of the present invention provides a transparent obstacle detection and map reconstruction system based on laser radar point cloud data, comprising: An information collection module, used to sense the target area using the laser radar of the unmanned vehicle; A computing module, which is used to store the data collected by the unmanned vehicle, deploy a recognition model for identifying transparent obstacles, identify the transparent obstacles from the collected data, and then output correction data; A map construction module is used to correct the data collected by the unmanned vehicle according to the correction data, and to construct a grid map of the target area accordingly; Communication module, which uses wired or wireless network to connect external devices for real-time status acquisition and remote control of unmanned vehicles; The power module is used to supply power to the information collection module, computing module, map building module and communication module of the unmanned vehicle.

[0020] Beneficial effects: The present invention provides a method and system for transparent obstacle detection and map reconstruction based on LiDAR point cloud data. First, the LiDAR point cloud data of glass in the surrounding environment is collected, and the processed point cloud data is used for deep learning model training; secondly, the point cloud data collected in real time is directly input into the model. If glass doors, windows or walls are detected, the deep learning model will obtain their location information; then, the point cloud data is interpolated and completed, and finally a reconstructed map with glass doors, windows or walls as wall obstacles is output. The present invention can solve the problem of the inability to correctly identify obstacles and affect the driving safety of unmanned vehicles in the existing mapping process due to the refraction and diffuse reflection of lasers by indoor glass doors, windows or walls. The deep learning model directly uses point cloud data as input, and has the advantages of low computational complexity and simple deployment. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1The present invention is a flow chart of a transparent obstacle detection and map reconstruction method based on lidar point cloud data.

[0022] Figure 2 This is a comparison diagram of point cloud data before and after modification in Example 1 of the present invention; Figure 2 (a) is the obstacle demonstration image before data modification; Figure 2 (b) is the obstacle demonstration image after data modification.

[0023] Figure 3 This is a comparison diagram of the effects before and after the map reconstruction in Example 1 of the present invention; Figure 3 (a) is the original grid map of the target area; Figure 3 (b) is the reconstructed raster map of the target area. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments, and they should not be understood as limitations on the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In the description of the present invention, it should be understood that the terms used are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0025] Combine the following Figure 1-Figure 3 The present invention describes a transparent obstacle detection and map reconstruction method and system based on laser radar point cloud data.

[0026] Embodiment 1: This embodiment provides a transparent obstacle detection and map reconstruction method based on laser radar point cloud data. The overall process of the method is as follows: Figure 1 As shown, including: S1. The unmanned vehicle uses laser radar to scan the entire target area including transparent obstacles to obtain multiple sets of point cloud data; S2, using a registration algorithm to register the acquired multiple sets of point cloud data to the same coordinate system to form a point cloud data set; S3, dividing the point cloud data set by dynamic window sampling, and performing geometric structure analysis to find point cloud data that meets the characteristics of transparent obstacles; S4. Construct and train a recognition model for identifying transparent obstacles, and use the recognition model to identify the point cloud data that meets the characteristics of transparent obstacles obtained in step S3. If a part is identified as having a transparent obstacle, the point cloud density is supplemented in the output data to be equivalent to that of an ordinary obstacle, and the part is regarded as an ordinary obstacle; if a part is identified as not having a transparent obstacle, the original sparse point cloud is removed from the output data, and the part is regarded as having no obstacle; S5. Replace the corresponding original data in the original point cloud data set with the modified output data of step S4, and construct a grid map of the target area based on the modified output data.

[0027] The above steps are explained in detail below through specific examples.

[0028] In step S1, any path planning algorithm may be used first so that the driving route of the driverless vehicle can cover the entire indoor space. The driverless vehicle uses a laser radar to collect data from indoor areas containing transparent obstacles such as glass doors and windows or walls.

[0029] In this embodiment, the unmanned vehicle uses the A* algorithm for path planning and carries a Ydlidar-X3 laser radar device for data collection. The point cloud data obtained by the laser radar device consists of a frame header, laser radar scanning parameters, scanning distance information, and point cloud data intensity information. The scanning parameters are set as follows: the scanning start angle starts at +180° and ends at -180°, the angle between two adjacent scanning points increases by 0.5°, and the scanning time increases by 0.0001 seconds. The minimum distance that the radar can detect is 0.07 meters to a maximum distance of 16 meters; in the scanning distance information, since it is collected every 0.5° and the scanning range is 360°, the distance information is organized into a one-dimensional array containing 720 elements, each element representing the distance from the center of the radar to the detected object at the corresponding angle (in meters).

[0030] The purpose of step S2 is to register the point cloud data.

[0031] In the point cloud data obtained by each scan of the laser radar, the origin of the point cloud coordinates is the laser radar position, resulting in the point cloud data obtained each time being not in the same coordinate system. At the same time, since the unmanned vehicle may still be in a driving state, the point cloud data obtained under different scanning angles of the laser radar may also not be in the same coordinate system. Therefore, it is necessary to perform a registration operation on the point cloud data to obtain a point cloud data set that can describe the complete structural information of the room in a unified coordinate system. The registration algorithms that can be used include but are not limited to the iterative closest point algorithm, the normal distribution transformation algorithm, etc., which are not limited in this embodiment.

[0032] Step S3 is to perform geometric structure analysis on the registered point cloud data set by dynamic window sampling to find the part that meets the characteristics of transparent obstacles.

[0033] First, the characteristic geometric features of the transparent obstacle point cloud are determined as follows: the point clouds at the front and back ends of the data are relatively dense, which is reflected as the edge of the transparent obstacle, and there are fewer and discontinuous point clouds in the middle, which is reflected as a transparent obstacle. If the number of point clouds in the straight line area fitted by the two edge coordinates is less than a given threshold , then it is considered that this part of the data may be a transparent obstacle. The threshold here It can be obtained through data pre-training.

[0034] Secondly, this embodiment uses dynamic window sampling to sample the point cloud data set. The specific method is: S31, setting the size of the initial window, and traversing the point cloud data set with the initial window; S32, performing geometric structure analysis on the point cloud data in the current window to find out the point cloud data having geometric features of transparent obstacles; S33, increase the window size, and perform the next traversal of the point cloud data set until all suspected transparent obstacles of different sizes are completely detected.

[0035] This requires pre-setting the window size adjustment rule: each time the window completes sampling of the point cloud data set, the density gradient between adjacent sampling windows is calculated , density gradient The calculation method is as follows: the number of point clouds within each sampling window is counted as the local density of the window, and the density gradient is obtained by calculating the density difference between adjacent windows.

[0036] If the window is first sampled or the gradient density changes by more than the feature loss threshold , then through the magnification factor To expand the window, the magnification factor The calculation formula is:

[0037] in, It is an adjustable coefficient, which is used to adjust the weight of gradient density in the formula. The adjustment coefficient needs to be set according to the experimental scenario; The calculated magnification factor , set the new window size , The calculation formula is:

[0038] In the formula, The original window size.

[0039] The purpose of step S4 is to further determine whether the point cloud data contains transparent obstacles through the neural network, and then modify the network output.

[0040] This step first requires building a recognition model for identifying transparent obstacles based on a neural network as a neural network model and training it.

[0041] The architecture of the recognition model is: The input layer obtains the data input from the sampling window and processes it into a tensor format before passing it to the feature extraction layer; The feature extraction layer first uses a multi-layer perceptron to independently extract features from each point cloud tensor, then obtains a global feature vector of fixed length through maximum pooling and fuses it with local features. The output dimension is The characteristic matrix of Represents the number of point clouds input into the network for each window, Indicates the dimension of the coordinate system.

[0042] The output layer takes the feature matrix as input and outputs the probability of containing transparent obstacles in the point cloud after processing by the fully connected layer and the softmax function. If the probability exceeds the threshold, it is determined that there is a point cloud with transparent glass obstacles. The probability threshold needs to be set according to the test scenario.

[0043] The various parameters of the recognition network need to be obtained after network training. The point cloud data set detected by the radar can be annotated to generate a training set, and the recognition model can be trained multiple times with the training set to adjust the model parameters to the optimal values ​​to obtain the recognition model.

[0044] In this embodiment, the data collection scene is a laboratory indoor scene, which contains transparent glass doors of different lengths and sizes. The unmanned vehicle is controlled to drive in the scene and collect the point cloud data output by the laser radar. In this embodiment, 360 sets of point cloud data are collected as training data for the recognition network, which are divided into training set and validation set in a ratio of 7:3. The optimizer used for training the network is AdamW, the learning rate adjustment method is the cosine annealing algorithm, the loss function of the training process is the cross entropy loss and the Dice loss, and the training round is set to Epochs, and the Epochs range is between 100 and 300.

[0045] The data dimension of the sampling window input recognition network is ( , 2), Indicates the number of point clouds input into the network in each window, and 2 indicates that it contains two-dimensional coordinate position information; in the feature extraction layer, the data is upgraded to ( , 128), that is, the original two-dimensional features of the point cloud are expanded to 128-dimensional features to highlight the information characteristics of the point cloud; the data after dimensionality increase is input into the maximum pooling layer, and after maximum pooling, data of size (1, 128) is obtained; finally, the data is input into the output layer, and in the output layer, the data of size (1, 2) is obtained through the fully connected layer, where 2 represents the prediction score of whether there is a transparent obstacle, and then the two scores are converted into probability distribution through the softmax function, and each value represents the probability that the point cloud belongs to the corresponding category.

[0046] After the model training is completed, the recognition model is used to recognize the point cloud data that meets the characteristics of transparent obstacles obtained in step S3. Before the data is input into the recognition model, the window sampling data is preprocessed. The purpose of the preprocessing is to make the data length in the input recognition network consistent for easy processing. The preprocessing algorithm includes but is not limited to the farthest point sampling algorithm, the average sampling algorithm, etc. In this embodiment, the preprocessing algorithm for the point cloud adopts the farthest point sampling, and the sampling distance threshold in the farthest point sampling is 0.3 meters.

[0047] Identify the probability of whether the point cloud in the model output window contains transparent obstacles. If it exceeds the set probability threshold , it is determined that this part of the point cloud contains transparent obstacles, and the two-dimensional coordinate information of the edge of the transparent obstacle in the map is output. In the embodiment, the threshold is set to 0.6.

[0048] For point clouds containing transparent obstacles, the missing point clouds are interpolated and completed according to the edge information of the transparent obstacle point clouds, and they are regarded as ordinary obstacles. In this embodiment, linear interpolation is used to complete the point cloud information with edge coordinate information as input. For point clouds without transparent obstacles, the original sparse point clouds are removed and they are regarded as no obstacles. The difference before and after data modification is as follows: Figure 2 As shown, Figure 2 (a) is the obstacle demonstration diagram before data modification. Figure 2 (b) is a demonstration image of obstacles after data modification. Comparing the two images, it is obvious that transparent obstacles have been completed and non-transparent obstacles have been deleted.

[0049] The modified data is transformed in the reverse processing mode of the preprocessing, so that the length of the output data is restored to the normal data length, which is convenient for filling into the original point cloud data.

[0050] Step S5 is a map reconstruction step.

[0051] The output data of the recognition model is used to replace the corresponding original data in the original point cloud data set to obtain the corrected target area grid map. The maps constructed before and after data correction are as follows: Figure 3 As shown, Figure 3 (a) is the original grid map of the target area; Figure 2 (b) is the reconstructed grid map of the target area. When the unmanned vehicle plans the path according to the corrected map, it can accurately avoid glass obstacles.

[0052] In addition, during the subsequent operation of the unmanned vehicle, the lidar continues to work. If there is a dynamic change in the opening or closing state of the glass door, the above steps are repeated to continuously update the grid map of the target area.

[0053] Embodiment 2: This embodiment provides a transparent obstacle detection and map reconstruction system based on laser radar point cloud data. The system is deployed on an unmanned vehicle as a physical carrier for executing the method of embodiment 1. It includes: The information collection module is used to sense the target area using the laser radar of the unmanned vehicle; correspondingly, step S1 in Example 1 is executed. It should be noted that this embodiment mainly uses a single-line laser radar to collect data, but is not limited to adding other sensors such as IMU sensors and GPS modules for information fusion processing.

[0054] The computing module is used to store the data collected by the unmanned vehicle, deploy a recognition model for identifying transparent obstacles, identify the transparent obstacles from the collected data, and then output correction data; correspondingly, steps S2 to S4 in Example 1 are executed.

[0055] The map construction module is used to correct the data collected by the unmanned vehicle according to the correction data, and construct a grid map of the target area accordingly; correspondingly, step S5 in Example 1 is executed.

[0056] The communication module uses a wired or wireless network to connect to external devices for real-time status acquisition and remote control of unmanned vehicles.

[0057] The power module is used to supply power to the information collection module, computing module, map building module and communication module of the unmanned vehicle.

[0058] The system embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0059] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A transparent obstacle detection and map reconstruction method based on laser radar point cloud data, characterized in that: include: S1. The unmanned vehicle uses laser radar to scan the entire target area including transparent obstacles to obtain multiple sets of point cloud data; S2, using a registration algorithm to register the acquired multiple sets of point cloud data to the same coordinate system to form a point cloud data set; S3, dividing the point cloud data set by dynamic window sampling, and performing geometric structure analysis to find point cloud data that meets the characteristics of transparent obstacles; S4. Construct and train a recognition model for identifying transparent obstacles, and use the recognition model to identify the point cloud data that meets the characteristics of transparent obstacles obtained in step S3. If a part is identified as having a transparent obstacle, the point cloud density is supplemented in the output data to be equivalent to that of an ordinary obstacle, and the part is regarded as an ordinary obstacle; if a part is identified as not having a transparent obstacle, the original sparse point cloud is removed from the output data, and the part is regarded as having no obstacle; S5. Replace the corresponding original data in the original point cloud data set with the modified output data of step S4, and construct a grid map of the target area based on the modified output data.

2. The transparent obstacle detection and map reconstruction method based on laser radar point cloud data according to claim 1 is characterized in that: Step S3 divides the point cloud data set by dynamic window sampling, and performs geometric structure analysis to find point cloud data that meets the characteristics of transparent obstacles. The specific method is as follows: S31, setting the size of the initial window, and traversing the point cloud data set with the initial window; S32, performing geometric structure analysis on the point cloud data in the current window, and finding point cloud data with geometric features of transparent obstacles. The geometric features of transparent obstacles are as follows: the point clouds at the front and back ends of the data are relatively dense, which are reflected as the edges of transparent obstacles; there are fewer and discontinuous point clouds in the middle, which are reflected as transparent obstacles; if the number of point clouds in the straight line area fitted by the coordinates of the two edges is less than a given threshold, it is considered that the part of the point cloud data is suspected to be a transparent obstacle; S33, increase the window size, and perform the next traversal of the point cloud data set until all suspected transparent obstacles of different sizes are completely detected.

3. The transparent obstacle detection and map reconstruction method based on laser radar point cloud data according to claim 2 is characterized in that: In step S3, each time the window completes sampling of the point cloud data set, the density gradient between adjacent sampling windows is calculated. If the window is sampled for the first time or the gradient density changes exceed the feature loss threshold, the magnification factor To expand the window, the magnification factor The calculation formula is: ； in, is an adjustable coefficient used to adjust the weight of the gradient density in the formula; The calculated magnification factor , set the new window size , The calculation formula is: ； In the formula, The original window size.

4. The transparent obstacle detection and map reconstruction method based on laser radar point cloud data according to claim 3 is characterized in that: The density gradient The calculation method is as follows: the number of point clouds within each sampling window is counted as the local density of the window, and the density gradient is obtained by calculating the density difference between adjacent windows.

5. The transparent obstacle detection and map reconstruction method based on laser radar point cloud data according to claim 1, characterized in that: The recognition model for identifying transparent obstacles constructed in step S4 is a neural network model, including: The input layer obtains the data input from the sampling window and processes it into a tensor format before passing it to the feature extraction layer; The feature extraction layer first uses a multi-layer perceptron to independently extract features from each point cloud tensor, then obtains a global feature vector of fixed length through maximum pooling and fuses it with local features. The output dimension is The characteristic matrix of Represents the number of point clouds input into the network for each window, Indicates the dimension of the coordinate system; The output layer takes the feature matrix as input and outputs the probability of containing transparent obstacles in the point cloud after processing by the fully connected layer and the softmax function. If the probability exceeds the threshold, it is determined that there is a point cloud with transparent glass obstacles.

6. The transparent obstacle detection and map reconstruction method based on laser radar point cloud data according to claim 5 is characterized in that: The method for training the recognition model is: annotating the point cloud data set detected by the radar to generate a training set, training the recognition model multiple times with the training set, adjusting the model parameters to the optimal values, and the loss functions of the training process are cross entropy loss and Dice loss.

7. The transparent obstacle detection and map reconstruction method based on laser radar point cloud data according to claim 5, characterized in that: The data inputted by the sampling window in the input layer is preprocessed to a specified data length, and the preprocessing method is one of the farthest point sampling algorithm or the average sampling algorithm; When the output layer outputs data, it uses the reverse processing method of preprocessing to restore the length of the output data to the normal data length.

8. The transparent obstacle detection and map reconstruction method based on laser radar point cloud data according to claim 1, characterized in that: The specific method of step S1 is: S11. Planning a path for the unmanned vehicle so that the driving route of the unmanned vehicle can cover the entire target area; S12. The unmanned vehicle uses a laser radar to scan at a set scanning point, and obtains a set of point cloud data after each 360° scan; the point cloud data at least includes distance information and point cloud data intensity information, and the scanning resolution is determined by the radar model.

9. The transparent obstacle detection and map reconstruction method based on laser radar point cloud data according to claim 1, characterized in that: After constructing the grid map, if there is a dynamic change in the opening or closing state of the glass door, return to step S2, correct the point cloud data acquired in real time to the point cloud data set, re-identify the transparent obstacles, and update the target area grid map.

10. A transparent obstacle detection and map reconstruction system based on laser radar point cloud data using the method according to any one of claims 1 to 9, characterized in that: include: An information collection module, used to sense the target area using the laser radar of the unmanned vehicle; A computing module, which is used to store the data collected by the unmanned vehicle, deploy a recognition model for identifying transparent obstacles, identify the transparent obstacles from the collected data, and then output correction data; A map construction module is used to correct the data collected by the unmanned vehicle according to the correction data, and to construct a grid map of the target area accordingly; Communication module, which uses wired or wireless network to connect external devices for real-time status acquisition and remote control of unmanned vehicles; The power module is used to supply power to the information collection module, computing module, map building module and communication module of the unmanned vehicle.

Images