POS辅助航带间航摄影像的自动转点  

POS-supported Matching Method for Aerial Images between Neighboring Strips

作　　者：袁修孝[1] 明洋[1] 

机构地区：[1]武汉大学遥感信息工程学院,湖北武汉430079

出　　处：《测绘学报》2010年第2期156-161,共6页Acta Geodaetica et Cartographica Sinica

基　　金：国家自然科学基金(40771176);国家创新研究群体科学基金(40721001)

摘　　要：提出一种利用POS提供的影像定向参数辅助区域航空摄影测量中航带间影像自动转点方法。首先,对航带内影像采用金字塔影像匹配策略寻求同名像点,以获得航向间影像连接点的精确像片坐标;然后利用POS提供的影像定向参数预测航带间的影像连接点。再通过对相邻航线间待匹配窗口影像的几何变形改正,在由POS数据辅助建立的核线几何条件约束下,直接在原始影像上找寻同名像点;最后经最小二乘匹配使影像匹配结果达到子像素级精度。试验表明,将POS数据引入航带间的影像自动转点中,不但可以简化影像匹配过程、提高自动化水平,而且较常规影像匹配方法可以提高航带间自动转点的成功率34.0%,并节省76.7%的CPU耗时。This paper proposes a novel image matching method for the aerial images between strips by using the exterior orientation parameters obtained via a position and orientation system（POS）.Firstly,all images within a strip are matched on the pyramid images to obtain the precise image coordinates of image tie points.Secondly,the approximate positions of these tie points between neighboring strips are predicted through the image exterior orientation parameters obtained via the POS.Thirdly,the image geometric distortions to be matched window in neighboring strip are corrected,and the corresponding points are directly searched on original images under the epipolar geometrically constrained condition.Finally,least squares image matching methods is performed in order to refine the image matching results to the sub-pixel order accuracy.Experiments have shown that POS-supported image matching method not only simplifies the procedure of image matching and increases the level of automation,but also improves the image matching success rate by 34.0% and save the CPU consuming time by 76.7% compared with the conventional image tie point transfer method between neighboring strips.

关 键 词：航摄影像 POS系统 影像匹配 匹配成功率 匹配效率 

分 类 号：P231[天文地球—摄影测量与遥感]