Study on workflow of acquiring, processing, and classifying LiDAR point cloud for establishment of 3D city models

Abstract

Along with the development of mapping technologies, the data acquisition system through the aerial laser scanning system has created a huge LiDAR point cloud data source. This is an essential and detailed data source that effectively serves various fields such as making maps and 3D maps, etc. The process of building 3D city models requires processing many types of data, at which point cloud data processing and classification play an essential role in creating an input data source for the model. Study on workflow for acquiring, processing, and classifying point cloud data is meaningful. It plays a vital role in the development and application of LiDAR technique in the current period, especially in building smart cities. The article introduces a method for acquiring, processing, and classifying LiDAR point cloud data to establish 3D city models. Based on the proposed approach, a pilot 3D model of the city in Thuong Ly ward, Hong Bang district, Hai Phong city was generated. The result of point cloud classification’s accuracy is very good, ranging from 92% to 99% depend on type of object. Based on the classification, the city’s 3D model is established with 2nd Level of Detail (LOD2) for Thuong Ly ward. The model is very useful in urban planning and management, and in the number of managing task such as infrastructure development, rescues and disaster prevention, military services and real estate business.