基于SIFT特征点的无人机航拍苗圃图像拼接

UAV Aerial Nursery Image Mosaic Based on SIFT Feature Points

针对航拍无法直接获取完整苗圃图像信息的问题,以无人机航拍苗圃图像为研究对象,利用尺度不变特征变换(SIFT)算法进行图像配准和加权平均法进行图像融合完成图像拼接,并针对SIFT算法出现误匹配的问题,利用随机抽样一致性(RANSAC)算法进行匹配点提纯。经过实验发现:本算法耗时78.97s,拼接结果与理想结果相似度可达99%。

To the problem that aerial image cannot contain the complete information of nursery,this paper took UAV aerial nursery image as study object.In order to achieve the goal of image mosaic,SIFT and weighted average algorithms were utilized to proceed image registration and image fusion.Also,RANSAC algorithm was used to wipe off the wrong matching points.Through the experiment,it is shown that the whole mosaic procedure costs 78.97 s,and the similarity between the actual result and desired one can reach 99%.