An Expectation Maximization Algorithm for LiDAR Point Cloud Classification

Abstract

LiDAR (Light Detection and Ranging) technology is now commonly used in geospatial technology when it is an active remote sensing technology and capable of collecting data on large areas. However, with a large dataset of measurement areas, selecting and using the data in accordance with the research purpose takes a lot of time to conduct the classification of points. The algorithm selection in LiDAR data processing problem is important in the use of lidar data. EM (Expectation Maximization) algorithm is a typical algorithm of data mining, with the advantage of being easy to install and understand the algorithm used in classification problems. In this study, the author improved the EM algorithm in classification of lidar point cloud by using scheduling parameters to reduce the convergence time of the algorithm.