摘要
针对噪声环境下的图像拼接方法,本文给出了一种基于小波包变换与数学形态学的方法,首先是对拼接图像进行小波降噪,然后利用形态学寻找主要特征物进行定位和对比,再通过样条函数构造的小波包对图像在不同尺度下的小波包分量进行拼接。实验表明,该方法可以较好地消除拼缝的影响,得到很好的拼接图。
In this paper, we present a stitching algorithm for overlapping noise images based on wavelet packet transform and mathematical morphology. Firstly, removing noise by wavelet transform and extracting main characteristic component by mathmatical morphology, then stitching each scale images based on B-spline wavelet packet transform. Experimental results indicate the method makes images match correctly and stitch images seamless and smooth.
出处
《激光与红外》
CAS
CSCD
北大核心
2006年第3期238-240,共3页
Laser & Infrared
关键词
多尺度分析
小波包变换
形态学
样条插值
图像拼接
multi-scale
wavelet packet transform
mathematical morphology
B-spline
image stitch
