基于多项式变换整体平差的航空影像拼接  

Aerial Image Stitching Based on Integral Adjustment of Polynomial Transformation

作　　者：何江[1] 邓非[1] 李朋龙[1] 宋涛[1] 

机构地区：[1]武汉大学测绘学院,湖北武汉430079

出　　处：《测绘地理信息》2016年第2期40-43,共4页Journal of Geomatics

基　　金：国家科技支撑计划资助项目(2012BAJ23B03)~~

摘　　要：针对目前已有的航空影像拼接技术中,变换参数仅仅依据拼接顺序独立求解,而未考虑到待拼接影像的整体关系的问题,提出一种用于航空影像拼接的整体平差算法。对所有具有重叠区域的待拼接影像进行匹配,以多项式变换为模型,利用最小二乘原理整体解求待拼接影像的变换参数。实验结果表明,该算法能较好地解决拼接误差随拼接次数增多而累积的问题,提高影像拼接的精度。The disadvantage of existing aerial image stitching algorithms is that transformation parameters are solved by the stitching sequence without considering the integral relationship of all images.This paper proposes an integral adjustment algorithm for aerial image stitching,which matches all the image pairs with overlapping region,chooses polynomial transformation as model,and integrally solves transformation parameters of all images by least-squares method.The experimental results show that this algorithm can solve error accumulation with the increasing number of stitching times and improve the accuracy of image stitching.

关 键 词：影像拼接 整体平差 多项式变换 最小二乘 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]