期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2007
卷号：XXXVI-3/W49A
页码：135-140
出版社：Copernicus Publications
摘要：Terrestrial laser scanning is becoming a standard for 3D modeling of complex scenes. Results of the scan contain detailed geometric information about the scene; however, the lack of semantic details is still a gap in making this data useable for mapping. In this paper we propose a framework for object recognition in laser scans. The 3D point cloud, which is the natural representation of scanners outcome, is a complex data structure to process, as it does not have an inherent neighborhood structure. We propose a polar representation which facilitates low-level image processing tasks, e.g. segmentation and texture modeling. Using attributes of each segment a feature space analysis is used to classify segments into objects. This process is followed by a fine-tuning stage based on graph-cut algorithm, which takes into consideration the 3D nature of the data. The proposed algorithm is demonstrated on tree extraction and tested on 18 urban scans containing complex objects in addition to trees. The experiments show the feasibility of the proposed framework
关键词：Object Recognition; Feature Extraction; Terrestrial Laser Scanner; Point Cloud; Algorithms
