| dc.description.abstract | This thesis involves point cloud extraction and automatic modeling. In point cloud extraction, we face difficulties such as high-dimensional data, noise and missing data, and irregularity and disorder. In order to solve these problems, we use the method of machine learning to extract features from point cloud, and perform tasks such as classification, and segmentation on point cloud. Machines extract features from data through learning, enabling us to effectively process high-dimensional data.
In addition, this research proposes an automatic modeling technique that combines methods such as segmentation of point clouds and principal components analysis. By segmenting the point cloud, we can distinguish different parts of the point cloud for better subsequent modeling operations. Using the method of principal value analysis, we can extract and analyze the normal vector of the whole part, so as to obtain a more accurate 3D model. The combination of these technologies enables us to extract object information from point cloud data and automatically model, and finally generate a 3D model. | en_US |