摘要
利用航空影像和LiDAR数据各自的优势点而进行的融合为复杂建筑物三维重建提供了一种可靠的途径。利用LiDAR数据获取建筑物初始边界,在航空影像中提取出边界特征,从而对初始建筑物边界进行精化;针对少量错误和丢失的建筑物边界进行编辑,开发了相应的软件系统。对某研究区域进行实验,并从重建建筑物模型的正确性、完整性和精度三方面进行分析,证明了所提出算法的可行性和有效性。
The integration of stereo-aerial images and LiDAR data has provided a promising methodology for building model generation because of the advantages and disadvantages of them. This paper starts by generating rough building boundary lines from LiDAR data. The rough boundaries are refined by introducing stereo aerial imagery. This process includes： straight line de- tection from stereo aerial images, line matching, 3D line grouping, precise boundary selection, and closed polygon generation. The experimental results are analyzed and discussed, meanwhile, the performance of the experimental results is evaluated from three aspects, correctness, completeness, and accuracy of the generated building models.
出处
《计算机工程与应用》
CSCD
2013年第2期13-18,23,共7页
Computer Engineering and Applications
基金
国家自然科学基金(No.41101409)
教育部留学回国人员科研启动基金(No.4002-122010)
湖北省自然科学基金项目(No.2010CDB09905)
教育部博士点基金(No.2011014612)
关键词
航空影像
LIDAR
建筑物重建
stereo-aerial images
LiDAR
digital building model generation
