无人机航拍图像配准算法研究  

Research on UAV aerial image registration algorithm

作　　者：史俊莉[1] 

机构地区：[1]河南工业职业技术学院建筑工程系,河南南阳473000

出　　处：《郑州轻工业学院学报（自然科学版）》2015年第1期63-66,共4页Journal of Zhengzhou University of Light Industry:Natural Science

基　　金：国家自然科学基金项目(61271279)

摘　　要：针对无人机航拍图像对尺度变化不明显的问题,在经典SIFT特征匹配算法的基础上,提出了一种改进的CS-SIFT特征匹配算法.该算法通过建立S层金字塔,达到降低多尺度空间和减少特征点数量的目的.在特征向量的匹配中,利用准欧氏距离替代常用欧氏距离,并通过极限几何约束,消除部分错误配准点对,进一步提高特征匹配效率.Matlab仿真结果表明,改进后的算法具有较高的匹配精度和较少的匹配时间,适用于对实时性要求较高的无人机航拍系统.Aiming at the problem that UAV aerial image was not obvious to scale change,an improved CS-SIFT feature matching algorithm was presented based on the classical feature matching algorithm of SIFT （scale invariant feature transform ）.The algorithm established S-layer pyramid to reduce the multi-scale space and the number of feature points.In the matching feature vectors,quasi-Euclidean distance substitu-ted commonly used Euclidean distance and geometric constraints limitd by eliminating part of the registra-tion error points to further improve the efficiency of feature matching.The simulation results using Matlab language indicated that the improved algorithm had higher matching accuracy and needed less matching time and it was quite suitable for the UAV aerial system of high real-time demand.

关 键 词：无人机航拍图像 CS-SIFT 尺度空间 准欧氏距离 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]