基于Curvelet变换的自适应遥感图像融合算法

SELF-ADAPTIVE FUSION ALGORITHM FOR REMOTE SENSING IMAGES BASED ON CURVELET TRANSFORM

提出一种基于Curvelet变换的自适应遥感图像融合算法。首先对全色图像和多光谱图像图像做Curvelet变换,分解成低频系数和高频系数,分解后的低频系数选取区域能量的加权系数自适应融合规则,高低频系数采用以空间频率为度量标准结合自适应因子进行自适应融合规则,再对得到的低频和高频系数进行Curvelet反变换得到融合图像。实验结果表明该算法是一种有效可行遥感图像融合算法。

This paper proposes a self-adaptive fusion algorithm for remote sensing images using Curvelet transform.Firstly,the Curvelet transform is applied to panchromatic images and multispectral images,and then they are decomposd into low-frequency coefficients and high-frequency coefficients.For decomposed low-frequency coefficients,we select self-adaptive fusion rule of weight coefficient of regional energy,and for high-frequency coefficients,a rule is used which self-adaptively fuses in metrics standard of space frequency combining the adaptive factor.The derived low-frequency and high-frequency coefficients are then performed the reverse Curvelet transform to acquire fused images.Experimental results show that this method is an effective and feasible fusion algorithm for remote sensing images.