摘要
为提高遥感影像间自动匹配的精度,采用SIFT(scale invariant feature)算法进行特征提取以及初匹配,并利用RANSAC(random sample consensus)和FSC(fast sample consensus)算法进行局外点剔除,分析剩余点的数量和分布情况。经过实验对比,基于SIFT+FSC算法,可以获取足够数量且分布均匀的匹配点对,实现遥感图像自动化的精确匹配,为影像的批量自动校正处理提供了解决思路。
Firstly the thesis performs feature extraction and first-step image registration by using SIFT algorithm,and removing the outliers through FSC(Fast Sample Consensus)algorithm and computing the parameters and then performing image registration and fusion process.Through the experimental comparison,a precise automatic image registration can be performed.This method provides a solution to the images of the far-shore islands and overseas that are lacking ground control points.
作者
郭丽
焦红波
董少敏
GUO Li;JIAO Hongbo;DONG Shaomin(92556 Troops,Zhoushan 310000,China;National Marine Data and Information Service,Tiaiijin 300171,China)
出处
《海洋测绘》
CSCD
北大核心
2021年第2期49-53,共5页
Hydrographic Surveying and Charting
关键词
遥感影像处理
特征提取
图像匹配
SIFT算法
FSC算法
remote sensing image processing
feature extraction
image registration
SIFT algorithm
FSC algorithm
