多图无缝拼接的配准算法  

A Matching Algorithm to Splice Multiple Images Seamlessly

作　　者：黄大坤 陆冬良[2] 严志明[1] 张秀彬[2,1] 

机构地区：[1]江西渝州科技职业学院 [2]上海交通大学电子信息与电气工程学院

出　　处：《微型电脑应用》2014年第2期62-65,共4页Microcomputer Applications

摘　　要：针对多幅待拼接数字图像所存在的曝光程度差异、图像配准误差、被摄物本身形态变化以及光学成像过程所产生的畸变等原因都会出现明显的拼接缝隙现象,即,在视觉效果上能够明显看得出存在"接缝线"和清晰地分辨出处于"接缝线"两侧存在着明显色差或亮度的原始相邻两幅图像等技术问题,提出一种多图像无缝拼接的智能配准融合算法。通过建立图像非线性畸变矫正数学模型、图像数据坐标转换、图像畸变校正及其求逆运算、寻找相邻两幅图像上的公共特征点、图像配准与融合等过程,能够完善地解决大型宽幅(广角)平面图像自动无缝拼接的技术难题。Because the digital image to be present in splicing differences in the degree of exposure,the image registration error,morphological changes of the subject itself,and the optical distortion produced by the imaging process are obvious reasons such splicing gap phenomenon that the visual effect can obviously see out there ＂seams＂ and be clearly distinguished technical problems in the ＂seams＂ on both sides of the obvious color or brightness of the original image,so we propose a intelligent matching ＆ fusing algorithm to splice multiple images seamlessly.The algorithm consists of image distortion correction nonlinear mathematical model,the image data coordinate transformation,image distortion correction and its inverse operation,look for the two common feature points on the adjacent images,image matching ＆ fusing process,etc.to solve large-scale improvement width （wide Angle） plane images automatic seamless technical problems.

关 键 词：图像匹配 无缝拼接 配准融合 畸变矫正 坐标转换 公共特征点 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]