Automatic representation and reconstruction of DBM from LiDAR data using Recursive Minimum Bounding Rectangle

摘要: Abstract Three-dimensional building models are important for various applications, such as disaster management and urban planning. The development of laser scanning sensor technologies has resulted in many different approaches efficient model generation using LiDAR data. Despite this effort, these lacks economical reliable techniques that fully exploit the advantage Therefore, research aims to develop a framework fully-automated by integrating data-driven model-driven methods datasets. starts employing data detection approximate boundary determination. generated boundaries then integrated into model-based processing strategy because derived planes show irregular due nature point acquisition. focus is generating buildings with right-angled-corners, which can be described collection rectangles under assumption majority areas belong category. applying Minimum Bounding Rectangle (MBR) algorithm recursively, decomposed sets further processing. At same time, quality MBRs examined verify buildings, from generated, right-angled-corners. parameters define primitives adjusted through fitting procedure boundaries. level details final Digital Building Model based on number recursions during MBR processing, turn determined density. improves seen experimental results, depends average spacing. This thus develops an approach not only automates generation, but also achieves best accuracy while utilizing