计算机应用研究2011，Vol.28Issue(6)：2342-2345,4.DOI:10.3969/j.issn.1001-3695.2011.06.094
基于全局仿射变换估计的航拍视频校正
Global affine motion estimation for aerial video registration
摘要
Abstract
This paper proposed a fast feature-based global affine motion estimation method for aerial video registration and motion detection.To establish the correspondences, developed the improved Harris corner detector to extract and select stable corners from image sequences, and extracted the SURF descriptor to describe those corners.Ahead of utilizing the random sample consensus technique for motion parameters estimation, viewed the matched couples as vectors, and analyzed to filter out the obvious mismatches to improve the ratio of inliers.Experiments demonstrate that the presented method can robustly estimate the global affine motion parameters with high accuracy, and can meet the requirements of motion detection for mobile platform.关键词
航拍视频/全局运动估计/角点检测/视频校正/特征匹配Key words
aerial video/ global motion estimation/ corner detection/ video registration/ feature matching分类
信息技术与安全科学引用本文复制引用
郭江,申浩,李书晓,常红星..基于全局仿射变换估计的航拍视频校正[J].计算机应用研究,2011,28(6):2342-2345,4.基金项目
国家自然科学基金资助项目(61005028,61005067) （61005028,61005067）
