一种双尺度分解的高分辨率遥感影像融合方法

A Method for Fusion of High-Resolution Remote Sensing Images Based on Dual-Scale Decomposition

针对不同分辨率的遥感影像融合时存在鲁棒性差、算法复杂度高的问题,提出了一种双尺度分解的高分遥感影像融合方法。该方法首先采用高分1号、高分2号不同分辨率的遥感影像作为实验数据,通过YUV提取多光谱影像的灰度分量Y;其次利用联合引导滤波的滚动引导滤波的方法将影像分为结构层与能量层,在能量层和结构层设计能量属性和梯度能量有效的影像融合规则并进行融合得到Y´分量;最后通过YUV逆变换得到最终的融合影像,并与其他4种方法进行对比实验,实验证明该方法是一种可行的高分辨率遥感影像融合方法。

Aiming at the problems of poor robustness and high complexity of the algorithm in the fusion of remote sensing images with different resolutions,a dual-scale decomposition method for the fusion of high-resolution remote sensing images is proposed.First,the remote sensing images with different resolutions of GF-1 and GF-2 are used as experimental data,and the grayscale component Y of the multispectral image is extracted by YUV;Then,the rolling-guided filtering method of joint-guided filtering is used to divide the image into structural layer and energy layer,and the energy attributes and gradient energy effective image fusion rules are designed in the energy layer and structural layer to fuse to get the component Y´;Finally,the YUV is inverted to get the final fused image,and the effectiveness and superiority of this method are compared with other four methods.Experiments indicate that this method is a feasible method for high-resolution remote sensing image fusion.