Research On Feature Extraction Algorithm Based On 3D Point Cloud Model

The feature line extraction of 3D point cloud model is the basis of point cloud data processing,3d reconstruction and point cloud registration.The feature line extraction of point cloud can be used to describe the salient features of the model,the research on feature line extraction of 3D point cloud model is of great significance and value to many fields in our production and life.At the same time,there are many problems in the research of feature line extraction of 3D point cloud model: The influence on the precision of feature line extraction should be reduced in the process of feature line extraction of model with large amount of point cloud data and high noise.And improve the efficiency and accuracy of complex minutiae feature extraction.With this as the research focus,the method of feature line extraction of 3D point cloud model is mainly studied.In order to solve the problems of high time complexity and low efficiency in feature extraction,a feature extraction method based on normal vector and projection plane is proposed in this paper.Firstly,principal component analysis is used to calculate the normal vector,then KNNS is used to search the k adjacent points of the detection point,and the method of combining projection plane and normal vector is used to increase the number of effective feature points,secondly,k-means algorithm is used to cluster the normal vectors,which can effectively reduce the influence of noise on feature point extraction.The feature points are sorted by regularization,and the feature lines of point cloud model are fitted by cubic b-spline algorithm.The results and advantages of the proposed method in the process of feature line extraction are verified by experiments.