自由像片重叠区域的确定及拼接  

Determining and Mosaicing the Overlap Area of Free Images

作　　者：王悦 吴云东[2] 张魁珂 

机构地区：[1]65015部队 [2]解放军信息工程大学测绘学院

出　　处：《海洋测绘》2008年第5期65-68,共4页Hydrographic Surveying and Charting

基　　金：江苏省普通高校高新技术产业发展项目(JHB05-36)

摘　　要：在获得低空航空遥感数据之后,通常需要将所获取的大量影像数据在没有地面控制的条件下进行概略拼接。由于低空航空遥感数据存在较大的旋偏角,其重叠区域的确定不同于简单的重叠确定。运用计算机图形学的方法,提出了一种确定重叠区域的通用算法,并采用改进SZELISKI算法进行影像拼接。试验结果表明:提出的通用算法能很好的确定影像重叠区域,且适用于多张影像重叠;改进的SZELISKI算法,其运算结果更加符合人的视觉习惯,影像过渡自然,重影现象最小。After we get low altitude aerial remote sensing data, it is required to mosaic the images without ground control points usually. Because there is many factors that affect low altitude aerial remote sensing fiat, the images have big angle of deviation sometimes, it is different from simple longitudinal overlap. We solve this problem through computer graphics and determine the overlap area. As a result, the current arithmetic also may find overlap area of several images. When mosaicing, we improve SZELISKI arithmetic, the results become more suitable to human optical custom. The final low altitude aerial remote sensing image connected naturally, the area of ghosting is small.

关 键 词：多边形 重叠区域 图像配准 图像拼接 

分 类 号：P231.1[天文地球—摄影测量与遥感]