一种基于SIFT匹配的影像位姿信息优化方法

应用GPS和IMU结合的POS系统制作正射图的方法生产效率高,成图周期短,成图质量取决于POS数据.采用SIFT特征匹配算法以及RANSAC算法剔除误匹配点,应用摄影测量学经典原理,得到对应影像的位姿信息,实现对原始POS数据的修正,通过最终空三结果以及图像拼接效果对比,得出的优化结果位姿更平整,更接近原始效果,消除了基于原始POS系统出现的畸变扭曲.

Image Position Information Optimization Method Based on SIFT Matching

It works efficiently making Digital Orthophoto Map by POS system which combined GPS and IMU.The cycle of making graph is short and the graph quality depends on the POS data.In this paper,the false matching points is eliminated by the SIFT feature matching algorithm and RANSAC algorithm.The position and orientation information is obtained by the application of photogrammetry classical principle.The POS data is corrected and it will draw the optimization conclusion by the aerotriangulation results and the contrast of the effect of image stitching finally,the position and orientation was more smooth and closer to the original effect,the distorted distortion based on the original POS system was eliminated.