UAV Image Quick Mosaic Algorithm Based on BSIFT
Shi Jian-don · Geomatics & Spatial Information Technology · 2015
This paper proposed a method BSIFT to deal with the low image registration speed problem of the traditional SIFT algorithm. Because SIFT feature point descriptor has computation complex,too much storage space,matching time- consuming long faults,we use BRIEF to replace it. The BSIFT algorithm is introduced to extract the feature points of UAV images and match them,RANSAC is applied to remove the mismatched points and estimate the parameters of the transformation matrix. The experimental results show that the image registration time is reduced obviously and the stitching accuracy is also improved.