基于影像特征点、线整体匹配的DSM自动生成  

Automatic DSM generation of global matching based on image feature point and edge

作　　者：李健[1] 刘凤德[2] 赵利平[2] 

机构地区：[1]武汉大学遥感信息工程院,武汉430079 [2]中国测绘科学研究院四维远见公司,北京100039

出　　处：《测绘科学》2007年第4期96-97,共2页Science of Surveying and Mapping

摘　　要：DSM的自动提取是“数字摄影测量”的最重要技术之一,其关键就是全自动影像匹配。其主要难点表现在:纹理信息缺乏、重复纹理、深度断裂与阴影、较大的投影变形等。为更好地解决上述问题,本文提出了基于特征点匹配与边缘线匹配相结合的方法。实验表明无论对山区还是城区都取得了很好的效果,证实了该方法在实际应用中的有效性。DSM extraction is one of the most important techniques in photogrammetry and remote sensing. Full-automatic image matching is the key problem and has many difficulties: poor and repetitive texture, terrain discontinuity, shadow, occlusion and projective deformation etc. To resolve those problems mentioned above, in this paper the authors propose a global matching method based on combination of feature point and feature edge. The experiments indicate that good results have been obtained in both mountain and urban area and prove this way is validate in practice.

关 键 词：DSM自动提取 边缘提取 特征点匹配 边缘线匹配 整体匹配 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]