基于网格变形和余弦函数权重的图像拼接方法  

An Image Stitching Method Based on Grid Warping and Cosine Function Weight

作　　者：李加亮 蒋品群[1] 夏海英[1] LI Jialiang;JIANG Pinqun;XIA Haiying(College of Electronic Engineering, Guangxi Normal University, Guilin Guangxi 541004, China)

机构地区：[1]广西师范大学电子工程学院,广西桂林541004

出　　处：《广西师范大学学报（自然科学版）》2020年第4期42-53,共12页Journal of Guangxi Normal University:Natural Science Edition

基　　金：国家自然科学基金(61762014);桂林市科学研究与技术开发计划(20170113-4);广西研究生教育创新计划(XYCSZ2019075)。

摘　　要：为了消除拼接图像中的重影和色差缝隙问题,本文提出一种基于网格变形和余弦函数权重的图像拼接方法。首先使用尺度不变特征变换(scale-invariant feature transform,SIFT)算法和随机抽样一致(random sample consensus,RANSAC)算法对待拼接图像进行特征点的提取与匹配;接着利用高斯分布(Gaussian distribution,GD)模型去除错误匹配的特征点;然后采用余弦函数权重对逼近投影图像拼接模型(as-projective-as-possible image stitching,APAP)和全局相似变换模型进行加权融合;最后利用余弦函数权重对变形图像进行加权拼接。实验结果表明:本文方法的拼接效果优于对比的经典算法,能有效地消除拼接图像中的重影和色差缝隙。An image stitching method based on the grid warping and cosine function weight is proposed to eliminate the ghost and chromatic gap in the stitched image.Firstly,the SIFT(scale-invariant feature transform)and RANSAC(random sample consensus algorithms)are used for the extraction and matching of feature points in the image to be stitched.Secondly,the GD(Gaussian distribution)model is used to remove the mismatched feature points.Then the cosine function is used as the weighting function of APAP(as-projective-as-possible image stitching)and global similarity transformation model.Finally,the cosine function weight is used to weight the warped image.The experimental results show that the stitching effect of this method is better than that of the classic algorithm,and the method can effectively eliminate the ghost and chromatic gap in the stitched image.

关 键 词：图像拼接 图像变形 尺度不变特征变换 权重 重影 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]