1. INTRODUCTION
New technologies and the development of acquisition and processing hardware and software in geographic information are opening a new viewpoint for the management of geographical 3D data, making the world easier to understand. The different methods used to create 3D city models from remote sensing data can be categorized based on automation, based on input data or hybrid methods. The most used methods are automatic methods or semi-automatic, that use photogrammetry to extract 3D models. Then the LiDAR come to be a complement to the classics methods and ameliorate the constructed models accuracy. Nowadays, LiDAR is one the most used technology for accurate 3D urban reconstruction. Light Detection And Ranging or LiDAR is an active remote sensing system that uses laser energy. LiDAR data are spatially formed as point clouds data. Point clouds are large sets of 3D points define by their (x, y, z) coordinates. Therefore, the basic assumption used for modeling is to summarize a building by a set of plans and lines that compose it. The plans describe facades or roofs, while the straight lines represent the edges of facades or roofs as well.