图像的多尺度稀疏分解及其在遥感图像融合上的应用

Multi-scale Sparse Decomposition of Image and Its Application in Remote Sensing Image Fusion

提出了一种图像多尺度稀疏分解的新方法,联合局部离散余弦变换基和曲波变换基组成分解字典,通过控制字典系数从多个尺度把二维图像稀疏分解为纹理成分和卡通成分,并以此应用到遥感图像融合,提取有效尺度下高分辨率全色遥感图像的纹理成分和多光谱遥感图像的卡通成分,对二者进行稀疏重建得融合图像.实验结果表明,多尺度稀疏分解的遥感图像融合方法优于经典融合方法,融合结果具有更高的空间分辨率和更低的光谱失真,相比流行的稀疏重建法,该方法的执行速率得到大幅提升,且取得了更好的融合结果.

A new multi-scale sparse image decomposition method is presented, which combines local discrete cosine transform bases and curvelet transform bases to build the decomposition dictionary and controls the entries of the diction- ary to decompose the image into texture component and cartoon component. Then we apply this technique to the field of remote sensing image fusion. Via sparse decomposition, the effective scale texture component of high resolution RSI and cartoon component of multi-spectral RSI are selected to be fused er. The experiment results show that our method obtains higher spatial resolution and lower spectral distortion than classical fusion method, and achieves a higher algo- rithm speed and a better fusion result than popular sparse reconstruction fusion method.