一种基于多尺度稀疏分解的遥感图像融合新方法

A new method for remote sensing image fusion based on multi-scale sparse decomposition

为取得更好的遥感图像融合效果,结合形态成分分析的思想,提出了图像的多尺度稀疏分解方法。集合曲波变换基和局部离散余弦变换基组成分解字典,通过控制字典系数的大小,将二维图像从多个尺度稀疏分解为纹理成分和卡通成分;从图像融合的信息量角度出发,提出了基于多尺度稀疏分解的遥感图像融合方法,通过稀疏分解提取有效尺度下高空间分辨率图像纹理成分和多光谱图像卡通成分,并对二者进行稀疏重建得到融合图像。与已有的经典融合方法相比,该方法以较小的计算代价换取了更高的空间分辨率和更低的光谱失真;与稀疏重建法相比,该方法的执行速率有较大提升,且有更好的融合效果。因此,所提出的基于多尺度稀疏分解的遥感图像融合方法有一定的推广应用价值。

To achieve better effect of remote sensing image fusion,the authors put forward a multi-scale sparse image decomposition method based on morphological component analysis( MCA). It combines curvelet transform basis and local discrete cosine transform( DCT) basis to form the decomposition dictionary and controls the entries of the dictionary so as to decompose the image into texture component and cartoon component. From the aspect of the amount of information,a remote sensing image( RSI) fusion method based on multi-scale sparse decomposition was proposed. By using sparse decomposition,the effective scale texture component of high resolution RSI and cartoon component of multi-spectral RSI were selected to be fused together. Compared with the classical fusion methods,the proposed fusion method gets higher spatial resolution and lower spectral distortion with a little computation load.Compared with sparse reconstruction fusion method,it achieves a higher algorithm speed and a better fusion result.Therefore,the proposed image fusion method based on multi-scale sparse decomposition has certain application value.