摘要
针对目前立体影像匹配时普遍存在的配准稳定性与配准精度不能同时兼顾问题,该文采用金字塔和分块策略,首先在多尺度空间下进行粗匹配,获得大量的初始特征点集;然后通过限制金字塔影像的阶数对初始特征点进行坏点剔除和滤点均匀化来获得良好的初始特征点;最后通过进行分块影像并行化SIFT算法匹配来实现精确匹配,提高效率。实验结果证明,此遥感立体影像并行稳定匹配方法能够提高匹配的效率,对影像的旋转、缩放和视角变换具有更强的抵抗性,其配准精度较高,适用于大范围的遥感立体影像的匹配。
Aiming at the problem that the stability and the accuracy of stereo image matching is not currently widespread at the same time,this paper used pyramid and partition strategy.First of all,using the affine transformation of ellipse in the multi scale space to fit maximally stable extreme region,and the initial large feature points were obtained;then through the restriction orderp yramid image of the initial feature point of bad pixels,good initial feature points were obtained;finally,through the block image parallel SIFT algorithm to achieve accurate matching,improving efficiency.Experimental results showed that the remote sensing stereo images parallel matching method based on maximally stable extreme region could improve the matching efficiency,and the rotation,zoom and view of transformation of the images became more resistant;the matching precision could reach sub-pixel degree,which was applicable to a wide range of remote sensing images matching.
出处
《测绘科学》
CSCD
北大核心
2015年第12期44-49,共6页
Science of Surveying and Mapping
基金
National Natural Science Fund Project(No.41301386)
Open Research Fund of the Key Laboratory of China Academy Position Precision Navigation and Timing Technology(No.2012PNTT17)
