Precise point cloud segmentation method based on distance judgment function

Effective segmentation of point cloud data is an important step in point cloud processing, and it is also a popular research direction in 3D point cloud processing. Traditional region growing algorithms are simple and easy to implement, and are widely used in 3D point cloud segmentation. However, the disorder and complexity of point cloud data and the uncertainty of initial seed node selection lead to over-segmentation and under-segmentation. This paper proposes a region growing algorithm based on distance judgment function calculation. First, we use the octree method to establish the topological relationship of the point cloud data, and construct local k neighborhoods, and eliminate outliers based on its density information; second, we perform k neighborhood search on the data points to obtain the covariance matrix of the neighborhood points, and use principal component analysis to calculate the eigenvalues and eigenvectors of the matrix; we use the minimum spanning tree method to compare the vector dot product, adjust the direction of the normal vector, and ensure the global consistency of the point cloud data; through the average curvature and Gaussian curvature Combined with calculation, the minimum curvature point is selected as the initial seed node, which improves the stability of seed node selection and avoids repeated segmentation; introduces a distance judgment function to judge the attributes of the seed point, calculates the normal distance from the selected seed point to its tangent plane, and passes the distance Threshold divides the point cloud data into flat points and sharp points to improve the efficiency of point cloud adjustment; filter the neighboring points according to the angle between the normal of the seed point and the normal of the neighboring point; finally set the curvature threshold reasonably and determine Guidelines for regional growth. According to the experimental results of segmentation, the region growing algorithm based on the distance judgment function improves the accuracy and stability of part segmentation, and improves the quality of segmentation.