摘要
针对传统淡入淡出渐变因子图像拼接方法对重叠区域含有几何错位或复杂内容实现无缝拼接时所存在的局限性,论文提出了一种基于渐变因子的自适应图像拼接算法。首先依据图像邻域方差对噪声和模糊度的分布估计来构造融合加权因子;然后根据拼接图像在同一坐标系下的几何错位来确定正态渐变加权因子;其次基于融合加权和正态渐变加权因子构造空域自适应渐变因子,实现图像的自然过渡和无缝拼接。最后,对两种算法进行了比较试验研究。
Aimming to the limitations of the traditional fade in/out algorithm used to mosaic the image overlapped region with complex content and geometry alignment error,an algorithm based on the adaptive graduation fade factor is proposed in this paper.Firstly,a fused weighted factor is constructed with respect to the noise and fuzzy degree distribution variances of the neighboring image region.Secondly,a normal distribution graduation factor is fixed,used to solve the geometry alignment error problem.Thirdly,combing the fused weighted factor and the normal distribution graduation factor results in an adaptive graduation fade factor which can make the mosaic image smoothness and seamless.Finally,the comparative experiment research is performed on both of the algorithm.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第22期72-74,77,共4页
Computer Engineering and Applications
基金
河南省科技厅资助项目
河南省教委自然科学基金资助项目(编号:2003120015)
河南省高校创新人才工程项目
关键词
图像拼接
熵匹配
融合加权因子
正态渐变
自适应因子
image mosaics, entropy matching, fused weighted factor,normal distribution graduation,adaptive factor
