Realization And Technology Of Regional Buildings Reconstruction

Reconstruction of 3D building models are widely used in the fields such as digital city, real estate, urban planning and so on. It’s also an active research area. Traditional 3D building reconstruction, which is based on surveying and charting with models created by hands through 3D model software, is both time-consuming and labor-intensive.3D building reconstruction based solely on images is mainly based on theories including epipolar geometry constraint between images series and in most cases there is constraint with camera angles of the input images or manual help is needed to identify feature points in the images. Since remote surveying and charting is developing fast these years, LIDAR data has become one of the data resources during the research of 3D building models reconstruction. LIDAR contains height information of all objects on the ground, which indeed helps a lot in building extraction and model creation. Combining LIDAR data and images to create 3D building models has strong feasibility.In this thesis, we research regional building reconstruction and present a method for reconstruction of 3D building models from aerial images and LIDAR Data and then realize it. In the beginning we analyze the current research status in this field and summarize the major methods of 3D building model reconstruction. And then we introduce the method proposed by us. Two major steps of regional building reconstruction are discussed in detail—building edges extraction and 3D model creation. The thesis elaborates the way to create height-gray image based on LIDAR data, the way to extract buildings from it, and the way to coordinate all the information in VRML to create models. It presents a simple implementation system to reconstruct building models from LIDAR data and aerial images and in some extent solves the problem in building extraction caused by trees as to improve the accuracy building extraction.