一种基于统计模型的自适应窗口匹配方法  

A matching method with adaptive window based on a statistical model

作　　者：林宗坚[1,2] 刘正坤[1] 支晓栋[1] 

机构地区：[1]武汉大学遥感信息工程学院,武汉430079 [2]中国测绘科学研究院,北京100039

出　　处：《测绘科学》2010年第6期5-8,共4页Science of Surveying and Mapping

基　　金：国家科技支撑计划(编号:2006BAB15B0102);863重点项目(编号:2008AA121305-5)

摘　　要：提高匹配精度和效率是摄影测量的主要研究内容之一。本文基于统计模型计算窗口灰度波动及视差波动,根据视差和其不确定性估计调整窗口。在调整窗口的同时确定最佳匹配点位,实现动态窗口匹配,匹配精度达到子像素级。通过对无人机影像匹配的实验结果验证了模型的有效性和实用性。To improve the precision and efficiency of matching is one of the main research contents of photogrammetry.Based on a statistical model this paper calculated the fluctuations of window＇s disparity and intensity.Then,the window-sizes were adjusted according to the disparity and its uncertainty estimate.When the window was adjusted,the optimal matching point was ascertained.Hence,the dynamical window matching is realized and the precision of matching is sub-pixel.Experiments on Unmanned Air Vehicles＇image matching have validated the effectiveness and practicability of the model.

关 键 词：匹配窗口 自适应窗口匹配 视差增量 不确定性估计 

分 类 号：P236[天文地球—摄影测量与遥感]