PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
Contact-free measurement devices known as Terrestrial Laser Scanners (TLS) capture dense point clouds of objects or sceneries by obtaining the coordinates and intensity value of each individual point. The point clouds are noisy and dispersed. By converting "data" to "information", a mathematical surface approximation may effectively decrease data storage and organize point clouds without requiring direct manipulation of the data. Uses include conducting stringent statistical testing for deformation analysis in the context of monitoring landslides. Classification and segmentation algorithms may recognize and eliminate non-uniform features like trees and shrubs to provide a smooth and precise mathematical surface of the ground by reaching an ideal approximation. In order to lead the reader through the current techniques, we provide a comparison of approaches for classifying TLS point clouds. In addition to the conventional point cloud filtering techniques, we will examine machine learning classification algorithms that rely on the manual extraction of point cloud features and PointNet++, a deep learning strategy that uses automated feature extraction.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.