基于B样条梯度的融合图像算法  

Gradient Image Fusion Algorithm Based on B-Spline

作　　者：刘军华[1] 雷超阳[1] 

机构地区：[1]长沙通信职业技术学院计算机工程系,湖南长沙410015

出　　处：《吉首大学学报（自然科学版）》2012年第6期61-65,共5页Journal of Jishou University(Natural Sciences Edition)

基　　金：湖南省教育厅科学研究项目(10C720)

摘　　要：图像融合在目标识别、计算机视觉、医学、遥感图像处理等领域应用广泛.利用B样条梯度融合图像算法来解决图像融合,将融合的图像分成3×3子块,计算每个子块的B样条梯度.把待融合图像对应的子块梯度模值累加求和,计算每个子块梯度模值的权值并与每个子块相乘并求和,得到融合后的图像子块.模拟实验结果表明该方法是有效的,优于传统的主分量分析、小波变换和Laplacian金字塔等方法.Image fusion is widely used in target recognition, computer vision, medicine, remote sensing, im- age processing and many other fields. This paper proposes an image fusion method based on B-spline gra- dient. It first divides the source image into small blocks of 3 × 3 size, calculates separately the B-spline gradient for each sub-block,adds up the gradient modulus values,calculates the weight value of each sub- block gradient modulus value,multiplies each sub-block and respective weight value and obtains the fu- sion image sub-block. Simulation experiments and analysis show that this method is effective and has a better performance than traditional methods such as principal component analysis, wavelet transforma- tion and the Laplacian pyramid.

关 键 词：B样条 图像融合 熵 梯度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]