A laser point cloud registration method and system for autonomous driving

Abstract

The present invention relates to a laser point cloud registration method and system for autonomous driving. The method comprises the following steps: obtaining a source point cloud and a target point cloud captured by a laser radar in the autonomous driving, initializing noise weights based on density, and calculating fast point feature histograms of the source point cloud and the target point cloud; downsampling the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations, calculating the matching probability of source points in the source point cloud to target points in the target point cloud, counting source points with matching probabilities lower than a threshold as noise points, and calculating the noise weights for the next iteration using a dynamic noise estimation method based on the number of noise points; optimizing a transformation matrix and a translation vector based on the matching probability, determining the change in the transformation matrix and the translation vector relative to the last calculated transformation matrix and the translation vector, and stopping updating the transformation matrix and the translation vector when the change meets a preset condition; and transforming the source point cloud using the transformation matrix and the translation vector to complete the registration of the source point cloud and the target point cloud.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and in particular to a laser point cloud registration method and system for autonomous driving. Background Art

[0002] Autonomous driving technology not only improves traffic efficiency and reduces traffic accidents, but also frees drivers' hands, enhancing travel comfort and convenience. With the continuous advancement of autonomous driving technology, ensuring accurate environmental perception and decision-making in complex and changing environments has become a core issue in its implementation. In this regard, LiDAR, with its high-precision and long-range detection capabilities, has become a key tool for acquiring environmental information in autonomous driving systems. LiDAR generates high-density three-dimensional point cloud data, reflecting the geometric shape and spatial distribution of surrounding objects, helping autonomous driving systems perform perception tasks such as obstacle detection and path planning.

[0003] In autonomous driving, point cloud registration is a core technology for environmental perception and map updating. Point cloud registration aligns point cloud data from different viewpoints, at different times, or from different sensors to generate a unified 3D model or scene graph. LiDAR scans in various environments and conditions often contain various types of noise, such as sensor errors, interference from dynamic objects, and inconsistent environmental reflections. This noise can lead to inaccurate point cloud registration results, compromising the reliability and safety of autonomous driving systems. Summary of the Invention

[0004] To address the problem that point cloud registration in autonomous driving is easily affected by noise, resulting in inaccurate modeling or obstacle recognition, the present invention first provides a laser point cloud registration method for autonomous driving, the method comprising the following steps:

[0005] Obtain the source point cloud and target point cloud captured by the LiDAR in autonomous driving, initialize the noise weight based on the density of the source point cloud and target point cloud, and calculate the fast point feature histogram of the source point cloud and target point cloud;

[0006] The source point cloud and target point cloud are downsampled according to the fast point feature histogram and the number of iterations. The matching probability between the source point in the source point cloud and the target point in the target point cloud is calculated. The source points with matching probability lower than the threshold are counted as noise points. The noise weight of the next iteration is calculated using the dynamic noise estimation method based on the number of noise points.

[0007] Based on the matching probability, the transformation matrix and translation vector are optimized, and the change of the transformation matrix and translation vector relative to the last calculated transformation matrix and translation vector is determined. When the change meets the preset conditions, the transformation matrix and translation vector are stopped from being updated; the source point cloud is transformed using the transformation matrix and translation vector to complete the alignment of the source point cloud and the target point cloud.

[0008] Preferably, the noise weight is initialized based on the density of the source point cloud and the target point cloud, specifically:

[0009] For each point in the source point cloud and the target point cloud, calculate the density of points in its neighborhood;

[0010] After calculating the local density of each point in the source point cloud and the target point cloud, the density distribution is obtained, the mean of the density is obtained, and the noise weight is initialized using the mean of the density.

[0011] Preferably, downsampling the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations is specifically as follows:

[0012] Evaluate the importance score of each point based on its fast point feature histogram feature;

[0013] Sort the source point cloud and target point cloud according to the importance score of the points, obtain the number of iterations, and determine the proportion of points retained in this iteration based on the number of iterations;

[0014] The point with the highest importance score is selected from the sorted source point cloud according to the ratio as the point of the source point cloud, and the point with the highest importance score is selected from the sorted target point cloud according to the ratio as the point of the target point cloud.

[0015] Preferably, the noise weight of the next iteration is calculated using a dynamic noise estimation method according to the number of noise points, specifically:

[0016] Calculate the ratio of the number of noise points to the number of points in the source point cloud in this iteration;

[0017] Obtain the neighborhood of the source point in the original source point cloud, calculate the normal difference between the source point and the neighborhood, obtain an adjustment coefficient based on the normal difference, and obtain the noise weight of the source point in the next iterative calculation through the adjustment coefficient and the ratio.

[0018] Preferably, the adjustment coefficient is obtained according to the normal difference, specifically:

[0019] Calculate the average of the dot products of the unit normal vectors of the source point and each point in the neighborhood;

[0020] The adjustment factor is obtained by subtracting the average value from 1.

[0021] Preferably, the importance score of each point is evaluated based on the fast point feature histogram feature of each point, specifically:

[0022] Calculate the similarity of the fast point feature histogram of the source point and the target point. For each point in the source point cloud and the target point cloud, the maximum similarity is used as the score of the point.

[0023] The local density of the point in the original source point cloud or the original target point cloud is obtained, and the weighted value is obtained by using the local density to weight the score to obtain the importance score of the point.

[0024] Secondly, the present invention also provides a laser point cloud registration system for autonomous driving, which includes the following modules:

[0025] The initialization module is used to obtain the source point cloud and target point cloud captured by the lidar in autonomous driving, initialize the noise weight based on the density of the source point cloud and target point cloud, and calculate the fast point feature histogram of the source point cloud and target point cloud;

[0026] The downsampling module is used to downsample the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations, calculate the matching probability of the source point in the source point cloud to the target point in the target point cloud, count the source points with matching probability lower than the threshold as noise points, and calculate the noise weight of the next iteration using the dynamic noise estimation method according to the number of noise points;

[0027] The registration module is used to optimize the transformation matrix and translation vector based on the matching probability, determine the change of the transformation matrix and translation vector relative to the last calculated transformation matrix and translation vector, and stop updating the transformation matrix and translation vector when the change meets the preset conditions; use the transformation matrix and translation vector to transform the source point cloud to complete the registration of the source point cloud and the target point cloud.

[0028] Preferably, the noise weight is initialized based on the density of the source point cloud and the target point cloud, specifically:

[0029] For each point in the source point cloud and the target point cloud, calculate the density of points in its neighborhood;

[0030] After calculating the local density of each point in the source point cloud and the target point cloud, the density distribution is obtained, the mean of the density is obtained, and the noise weight is initialized using the mean of the density.

[0031] Preferably, downsampling the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations is specifically as follows:

[0032] Evaluate the importance score of each point based on its fast point feature histogram feature;

[0033] Sort the source point cloud and target point cloud according to the importance score of the points, obtain the number of iterations, and determine the proportion of points retained in this iteration based on the number of iterations;

[0034] The point with the highest importance score is selected from the sorted source point cloud according to the ratio as the point of the source point cloud, and the point with the highest importance score is selected from the sorted target point cloud according to the ratio as the point of the target point cloud.

[0035] Preferably, the noise weight of the next iteration is calculated using a dynamic noise estimation method according to the number of noise points, specifically:

[0036] Calculate the ratio of the number of noise points to the number of points in the source point cloud in this iteration;

[0037] Obtain the neighborhood of the source point in the original source point cloud, calculate the normal difference between the source point and the neighborhood, obtain an adjustment coefficient based on the normal difference, and obtain the noise weight of the source point in the next iterative calculation through the adjustment coefficient and the ratio.

[0038] Preferably, the adjustment coefficient is obtained according to the normal difference, specifically:

[0039] Calculate the average of the dot products of the unit normal vectors of the source point and each point in the neighborhood;

[0040] The adjustment factor is obtained by subtracting the average value from 1.

[0041] Preferably, the importance score of each point is evaluated based on the fast point feature histogram feature of each point, specifically:

[0042] Calculate the similarity of the fast point feature histogram of the source point and the target point. For each point in the source point cloud and the target point cloud, the maximum similarity is used as the score of the point.

[0043] The local density of the point in the original source point cloud or the original target point cloud is obtained, and the weighted value is obtained by using the local density to weight the score to obtain the importance score of the point.

[0044] In the initial stage, the present invention assigns lower noise weights to points in dense areas, which can more accurately distinguish between noise points and important points and improve the robustness of registration. Moreover, in the CPD iteration, combined with the matching probability and FPFH, the noise weights can be adjusted more flexibly to adapt to the dynamic changes of noise points during the point cloud registration process, making the registration results more accurate and stable, and suitable for complex autonomous driving scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flow chart of Example 1;

[0046] Figure 2This is a structural diagram of Example 2. DETAILED DESCRIPTION

[0047] In this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0049] In the first embodiment of the present invention, a laser point cloud registration method for autonomous driving is provided, such as Figure 1 As shown, the method includes the following steps:

[0050] S1, obtains the source point cloud and target point cloud captured by the lidar in autonomous driving, initializes the noise weight based on the density of the source point cloud and target point cloud, and calculates the fast point feature histogram of the source point cloud and target point cloud;

[0051] In autonomous driving, LiDAR continuously acquires point clouds around the vehicle. When registering a point cloud from one moment to the next, the previous point cloud is considered the target point cloud, while the next point cloud is considered the source point cloud. Alternatively, within a point cloud acquired at the same moment, one point cloud from different LiDAR sensors may be considered the source point cloud and the other the target point cloud.

[0052] The Coherent Point Drift (CPD) algorithm includes the E step and the M step. In the E step, the formula is used. Calculation Point The probability distribution of , where Represents the noise weight. A small noise weight may force noise points to participate in the matching, resulting in mismatches. A large noise weight may cause a large number of points to be considered as potential noise, which may weaken the geometric constraints on the real points and reduce the registration accuracy.

[0053] At the beginning of registration, noise weights are initialized based on the density of the point clouds (i.e., the distance between points or the distribution of points). For example, if the density of the source and target point clouds is low, it may mean that there are more noisy points, so these points will be given higher noise weights during initialization. In an optional embodiment, the noise weights are initialized based on the density of the source and target point clouds, specifically:

[0054] For each point in the source point cloud and the target point cloud, calculate the density of points in its neighborhood;

[0055] After calculating the local density of each point in the source point cloud and the target point cloud, the density distribution is obtained, the mean of the density is obtained, and the noise weight is initialized using the mean of the density.

[0056] For each point in the source point cloud and the target point cloud, the density of the points in its neighborhood is calculated. Density usually refers to the number of points within a certain radius around a point. Based on these density values, the local density distribution of each point is calculated, and the density mean of all points in the source point cloud and the target point cloud is obtained. The average value of the density is used to obtain the initialization noise weight. Specifically, the initialization noise weight is calculated using a negative exponential function or an inverse proportional function. If the point cloud is relatively sparse, then the density is low and the calculated noise weight is higher, which means that the point may contain more noise, otherwise the noise weight is lower. In one embodiment, the initialization noise weight of all points is the same.

[0057] Fast Point Feature Histograms (FPFH) are a method for describing the local geometric features of each point in a point cloud. They are particularly useful for identifying matching points in point cloud registration. The calculation process for FPFHs is well-known and will not be detailed here.

[0058] S2, downsample the source point cloud and target point cloud according to the fast point feature histogram and the number of iterations, calculate the matching probability of the source point in the source point cloud to the target point in the target point cloud, count the source points with matching probability lower than the threshold as noise points, and calculate the noise weight of the next iteration using the dynamic noise estimation method according to the number of noise points;

[0059] During the registration process, the FPFH feature can better match the points in the source point cloud and the target point cloud. In order to speed up the calculation and avoid processing too much data, the point cloud is downsampled according to the FPFH and the number of iterations, that is, the number of points in the point cloud is reduced. In one embodiment, the source point cloud and the target point cloud are downsampled according to the fast point feature histogram and the number of iterations, specifically:

[0060] Evaluate the importance score of each point based on its fast point feature histogram feature;

[0061] Sort the source point cloud and target point cloud according to the importance score of the points, obtain the number of iterations, and determine the proportion of points retained in this iteration based on the number of iterations;

[0062] The point with the highest importance score is selected from the sorted source point cloud according to the ratio as the point of the source point cloud, and the point with the highest importance score is selected from the sorted target point cloud according to the ratio as the point of the target point cloud.

[0063] Calculate the source point cloud and the target point cloud for each point

[0064] In step S1, the fast point feature histogram of each point in the source point cloud and the target point cloud is calculated, and the importance score of the point is evaluated according to the FPFH of each point in the source point cloud and the target point cloud. In a specific embodiment, the similarity of the fast point feature histograms of the source point and the target point is calculated, and for each point in the source point cloud and the target point cloud, the maximum similarity is used as the score of the point; the local density of the point in the original source point cloud or the original target point cloud is obtained, and the weighted value is obtained by using the local density. The score is weighted by the weighted value to obtain the importance score of the point.

[0065] Calculating the FPFH similarity between source and target points involves calculating the similarity between each point in the source point cloud and each point in the target point cloud, with the maximum similarity used as the point score. For example, if there are four points in the source point cloud and five points in the target point cloud, and the FPFH similarities between point 1 in the source point cloud and the five points in the target point cloud are 0.1, 0.2, 0.1, 0.3, and 0.5, respectively, then the maximum similarity for point 1 is 0.5. Similarly, the maximum similarity between each point in the target point cloud and the source point cloud is used as the score for the point in the target point cloud. The local density of the points in the original point cloud is then calculated, and a weighted value is obtained using the local density. In one embodiment, the greater the local density, the greater the weighted value. The weighted value is used to weight the scores to obtain the point importance score.

[0066] After obtaining the importance scores of the points, the source point cloud and the target point cloud are sorted respectively according to the importance scores of the points, the number of iterations is obtained, and the proportion of points retained in this iteration is determined according to the number of iterations. In one embodiment, the greater the number of iterations, the greater the proportion of retained points. For example, in the first iteration, the retained proportion is 20%, and in the second iteration, the retained proportion is 30%. According to the proportion, the point with the highest importance score is selected from the sorted source point cloud as the point of the source point cloud, and according to the proportion, the point with the highest importance score is selected from the sorted target point cloud as the point of the target point cloud; for example, if the retention ratio is 20%, the first 20% of the points are selected from the sorted source point cloud as the points of the downsampled source point cloud, and similarly, the first 20% of the points are selected from the sorted target point cloud as the points of the downsampled target point cloud. Obviously, the above sorting is sorted in descending order according to the importance score.

[0067] In the CPD algorithm, the matching probability is the matching situation between the source point and the target point, which is expressed by the formula Calculate the matching probability, where A is the transformation matrix, t is the translation vector, is the point of the source point cloud, is the point of the target point cloud, The source points with statistical matching probability lower than the threshold are regarded as noise points. That is, if the matching probability of a source point and all target points is less than the threshold, for example, 0.2, then the source point is regarded as a noise point.

[0068] In one embodiment, the noise weight of the next iteration is calculated using a dynamic noise estimation method according to the number of noise points, specifically:

[0069] Calculate the ratio of the number of noise points to the number of points in the source point cloud in this iteration;

[0070] Obtain the neighborhood of the source point in the original source point cloud, calculate the normal difference between the source point and the neighborhood, obtain an adjustment coefficient based on the normal difference, and obtain the noise weight of the source point in the next iterative calculation through the adjustment coefficient and the ratio.

[0071] The overall noise level is estimated by the proportion of noise points identified in the current iteration. An adjustment coefficient is calculated based on the normal difference between each point and its neighborhood. The adjustment coefficient is used to reflect the local geometric changes of the point. Points with larger differences are given higher noise weights, thereby reducing the dependence on these points in the next iterative calculation.

[0072] In one embodiment, obtaining the adjustment coefficient based on the normal difference is specifically as follows: calculating the average value of the dot product of the unit normal vectors of the source point and each point in the neighborhood; and subtracting the average value from 1 as the adjustment coefficient.

[0073] For example, in this iteration, the source point cloud has 100 points, 10 noise points are identified, and the noise ratio is 10%. For source point i, its neighborhood in the original source point cloud is found and the normal difference is calculated. For example, if the normal difference is 0.3 and the adjustment coefficient is 1-0.3 = 0.7, then in the next iteration, the noise weight of source point i = noise ratio × adjustment coefficient = 0.1 × 0.7 = 0.07.

[0074] S3, optimizes the transformation matrix and translation vector based on the matching probability, determines the change of the transformation matrix and translation vector relative to the last calculated transformation matrix and translation vector, and stops updating the transformation matrix and translation vector when the change meets the preset conditions; uses the transformation matrix and translation vector to transform the source point cloud to complete the alignment of the source point cloud and the target point cloud.

[0075] The CPD algorithm is a process of continuously iteratively updating the transformation matrix and translation vector. After each iteration, the change in the transformation matrix and translation vector relative to the previously calculated transformation matrix and translation vector is determined. When the change meets a preset condition, the update of the transformation matrix and translation vector is stopped. The transformation matrix and translation vector are then used to transform the source point cloud to complete the registration of the source and target point clouds. Optimizing the transformation matrix and translation vector based on matching probability is a step in the CPD algorithm and is a prior art technique, so it will not be further described here.

[0076] In a first embodiment of the present invention, a laser point cloud registration system for autonomous driving is provided. Figure 2 As shown, the system includes the following modules:

[0077] The initialization module is used to obtain the source point cloud and target point cloud captured by the lidar in autonomous driving, initialize the noise weight based on the density of the source point cloud and target point cloud, and calculate the fast point feature histogram of the source point cloud and target point cloud;

[0078] The downsampling module is used to downsample the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations, calculate the matching probability of the source point in the source point cloud to the target point in the target point cloud, count the source points with matching probability lower than the threshold as noise points, and calculate the noise weight of the next iteration using the dynamic noise estimation method according to the number of noise points;

[0079] The registration module is used to optimize the transformation matrix and translation vector based on the matching probability, determine the change of the transformation matrix and translation vector relative to the last calculated transformation matrix and translation vector, and stop updating the transformation matrix and translation vector when the change meets the preset conditions; use the transformation matrix and translation vector to transform the source point cloud to complete the registration of the source point cloud and the target point cloud.

[0080] Preferably, the noise weight is initialized based on the density of the source point cloud and the target point cloud, specifically:

[0081] For each point in the source point cloud and the target point cloud, calculate the density of points in its neighborhood;

[0082] After calculating the local density of each point in the source point cloud and the target point cloud, the density distribution is obtained, the mean of the density is obtained, and the noise weight is initialized using the mean of the density.

[0083] Preferably, downsampling the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations is specifically as follows:

[0084] Evaluate the importance score of each point based on its fast point feature histogram feature;

[0085] Sort the source point cloud and target point cloud according to the importance score of the points, obtain the number of iterations, and determine the proportion of points retained in this iteration based on the number of iterations;

[0086] The point with the highest importance score is selected from the sorted source point cloud according to the ratio as the point of the source point cloud, and the point with the highest importance score is selected from the sorted target point cloud according to the ratio as the point of the target point cloud.

[0087] Preferably, the noise weight of the next iteration is calculated using a dynamic noise estimation method according to the number of noise points, specifically:

[0088] Calculate the ratio of the number of noise points to the number of points in the source point cloud in this iteration;

[0089] Obtain the neighborhood of the source point in the original source point cloud, calculate the normal difference between the source point and the neighborhood, obtain an adjustment coefficient based on the normal difference, and obtain the noise weight of the source point in the next iterative calculation through the adjustment coefficient and the ratio.

[0090] Preferably, the adjustment coefficient is obtained according to the normal difference, specifically:

[0091] Calculate the average of the dot products of the unit normal vectors of the source point and each point in the neighborhood;

[0092] The adjustment factor is obtained by subtracting the average value from 1.

[0093] Preferably, the importance score of each point is evaluated based on the fast point feature histogram feature of each point, specifically:

[0094] Calculate the similarity of the fast point feature histogram of the source point and the target point. For each point in the source point cloud and the target point cloud, the maximum similarity is used as the score of the point.

[0095] The local density of the point in the original source point cloud or the original target point cloud is obtained, and the weighted value is obtained by using the local density to weight the score to obtain the importance score of the point.

[0096] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using a general-purpose hardware platform, or alternatively, through a combination of hardware and software. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a computer product. The present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0097] 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, and other embodiments may also be used. Although the present invention has been described in detail with reference to the aforementioned embodiments, ordinary technicians in this field should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A laser point cloud registration method for autonomous driving, characterized in that: The method comprises the following steps: Obtain the source point cloud and target point cloud captured by the LiDAR in autonomous driving, initialize the noise weight based on the density of the source point cloud and target point cloud, and calculate the fast point feature histogram of the source point cloud and target point cloud; The source point cloud and target point cloud are downsampled according to the fast point feature histogram and the number of iterations. The matching probability between the source point in the source point cloud and the target point in the target point cloud is calculated. The source points with matching probability lower than the threshold are counted as noise points. The noise weight of the next iteration is calculated using the dynamic noise estimation method based on the number of noise points. Based on the matching probability, the transformation matrix and translation vector are optimized, and the change of the transformation matrix and translation vector relative to the last calculated transformation matrix and translation vector is determined. When the change meets the preset conditions, the transformation matrix and translation vector are stopped from being updated. The source point cloud is transformed using the transformation matrix and translation vector to complete the registration of the source point cloud and the target point cloud. The noise weight is initialized based on the density of the source point cloud and the target point cloud, specifically: For each point in the source point cloud and the target point cloud, calculate the density of points in its neighborhood; After calculating the local density of each point in the source point cloud and the target point cloud, the density distribution is obtained, the mean of the density is obtained, and the initialization noise weight is obtained using the mean of the density; The downsampling of the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations is specifically as follows: Evaluate the importance score of each point based on its fast point feature histogram feature; Sort the source point cloud and target point cloud according to the importance score of the points, obtain the number of iterations, and determine the proportion of points retained in this iteration based on the number of iterations; Selecting the point with the highest importance score from the sorted source point cloud according to the ratio as a point of the source point cloud, and selecting the point with the highest importance score from the sorted target point cloud according to the ratio as a point of the target point cloud; The noise weight of the next iteration is calculated using a dynamic noise estimation method according to the number of noise points, specifically: Calculate the ratio of the number of noise points to the number of points in the source point cloud in this iteration; Obtain the neighborhood of the source point in the original source point cloud, calculate the normal difference between the source point and the neighborhood, obtain an adjustment coefficient based on the normal difference, and obtain the noise weight of the source point in the next iterative calculation through the adjustment coefficient and the ratio.

2. The method according to claim 1, wherein The adjustment coefficient is obtained according to the normal difference, specifically: Calculate the average of the dot products of the unit normal vectors of the source point and each point in the neighborhood; The adjustment factor is obtained by subtracting the average value from 1.

3. The method according to claim 1, wherein The importance score of each point is evaluated based on the fast point feature histogram feature of each point, specifically: Calculate the similarity of the fast point feature histogram of the source point and the target point. For each point in the source point cloud and the target point cloud, the maximum similarity is used as the score of the point. The local density of the point in the original source point cloud or the original target point cloud is obtained, and the weighted value is obtained by using the local density to weight the score to obtain the importance score of the point.

4. A laser point cloud registration system for autonomous driving, characterized in that: The system includes the following modules: The initialization module is used to obtain the source point cloud and target point cloud captured by the lidar in autonomous driving, initialize the noise weight based on the density of the source point cloud and target point cloud, and calculate the fast point feature histogram of the source point cloud and target point cloud; The downsampling module is used to downsample the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations, calculate the matching probability of the source point in the source point cloud to the target point in the target point cloud, count the source points with matching probability lower than the threshold as noise points, and calculate the noise weight of the next iteration using the dynamic noise estimation method according to the number of noise points; The registration module is used to optimize the transformation matrix and translation vector based on the matching probability, determine the change of the transformation matrix and translation vector relative to the last calculated transformation matrix and translation vector, and stop updating the transformation matrix and translation vector when the change meets the preset conditions; use the transformation matrix and translation vector to transform the source point cloud to complete the registration of the source point cloud and the target point cloud; The noise weight is initialized based on the density of the source point cloud and the target point cloud, specifically: For each point in the source point cloud and the target point cloud, calculate the density of points in its neighborhood; After calculating the local density of each point in the source point cloud and the target point cloud, the density distribution is obtained, the mean of the density is obtained, and the initialization noise weight is obtained using the mean of the density; The downsampling of the source point cloud and the target point cloud according to the fast point feature histogram and the number of iterations is specifically as follows: Evaluate the importance score of each point based on its fast point feature histogram feature; Sort the source point cloud and target point cloud according to the importance score of the points, obtain the number of iterations, and determine the proportion of points retained in this iteration based on the number of iterations; Selecting the point with the highest importance score from the sorted source point cloud according to the ratio as a point of the source point cloud, and selecting the point with the highest importance score from the sorted target point cloud according to the ratio as a point of the target point cloud; The noise weight of the next iteration is calculated using a dynamic noise estimation method according to the number of noise points, specifically: Calculate the ratio of the number of noise points to the number of points in the source point cloud in this iteration; Obtain the neighborhood of the source point in the original source point cloud, calculate the normal difference between the source point and the neighborhood, obtain an adjustment coefficient based on the normal difference, and obtain the noise weight of the source point in the next iterative calculation through the adjustment coefficient and the ratio.

Images