基于非完全拓扑分析的三维建筑重建识别算法  

Image-Based 3D Building Reconstruction based on Non-complete topological analysis

作　　者：王传安[1] 葛华[1] 王亚军[1] 

机构地区：[1]安徽科技学院,安徽凤阳233100

出　　处：《武汉工程大学学报》2011年第10期100-102,共3页Journal of Wuhan Institute of Technology

基　　金：应用型本科计算机公共基础课程群研究与实践(X201013);安徽省自然科学研究项目(KJ2011Z070)

摘　　要：为实现基于单幅图像的三维建筑物重建识别过程,本文提出了一种有效的重建方法———非完全拓扑分析.首先,对三维建筑物可见部分的轮廓进行拓扑分析并提取,得到完整的拓扑结构.其次,将分割后的拓扑结构与数据库中原始信息进行快速识别匹配.实验表明,本方法与SSDA图像匹配算法相比,大大提高了计算速度和匹配效率.Image-Based 3D Building Reconstruction is a very difficult problem in computer vision and computer graphics.An efficient method of 3D building reconstruction method is proposed by using non-complete topological analysis,which rapidly achieves the process of recognizing architectures based one or two images.Firstly,the outline of visible parts of 3D reconstruction is extracted from image and is analyzed to get the incomplete topology of architectures.Then,the resulted topology will be divided into some primitive geometry,and these geometries are used to match models in databases finally.Experiments show that this method VS SSDA has better effectiveness.

关 键 词：拓扑分析 三维重建 噪声阈值 识别算法 

分 类 号：TQ177.5[化学工程—硅酸盐工业]