基于边缘检测小波变换的图像融合研究

本文用基于极大值图像边缘检测的小波变换进行图像融合,即通过用B样条函数构造适用于图像边缘检测的小波对输入图像进行两个互相垂直方向的分解,然后分别用小波系数比较的融合准则和边缘比较的融合准则进行融合,实验发现由此融合的图像质量优于基于典型小波分解的融合.进行了仅用图像边缘融合并通过简单重构生成输出图像的实验,发现由此得到的融合图像质量可以在可接受的范围内,但可使参与融合的像素大幅度减少,这对于某些要求数据压缩,存储量较少的工程化图像融合系统有一定参考价值.

Study on Image Fusion Based on Wavelet of Edge Detection

The paper uses a wavelet of edge detection on image fusion.The wavelet constructed by B-spline function is appropriate for edge detection and used for image decomposition in two vertical directions.And then fuse image by the rule which compares coefficients of waveltet transform or compares edge of images.The result of experiment shows the quality of fusion image based on wavelet of edge detection precedes typical wavelets.Also try using edge's pixels only to fuse image and find the quality of fusion image is acceptable,but the number of joining pixels reduces greatly.This method may apply to some image fusion that requires data compression or less storage.