HCT变换与联合稀疏模型相结合的遥感影像融合

提出了一种基于HCT变换和联合稀疏模型的遥感影像融合方法,可更有效地利用多光谱所需谱段的光谱信息,最终得到所需谱段的融合影像。该方法将所需谱段的多光谱影像进行HCT变换,获取其亮度分量和角度分量;然后利用亮度分量和全色影像小波变换的低频分量进行联合稀疏模型的构建、系数求解和融合,得到融合的全色低频分量;最后将该低频分量与前面步骤所得其他分量分别进行小波逆变换和HCT逆变换,得到高质量的融合影像。试验利用Pleiades-1和WorldView-2两种卫星数据进行验证,并通过视觉效果和量化的融合评价指标进行对比和分析,验证了本文算法的有效性。

A Pansharpening Method Based on HCT and Joint Sparse Model

A novel fusion method based on the hyperspherical color transformation (HCT) and joint sparsity model is proposed for decreasing the spectral distortion of fused image further. In the method, an intensity component and angles of each band of the multispectral image is obtained by HCT firstly, and then the intensity component is fused with the panchromatic image through wavelet transform and joint sparsity model. In the joint sparsity model, the redundant and complement information of the different images can be efficiently extracted and employed to yield the high quality results. Finally, the fused multi spectral image is obtained by inverse transforms of wavelet and HCT on the new lower frequency image and the angle components, respectively. Experimental results on Pleiades-1 and WorldView-2 satellites indicate that the proposed method achieves remarkable results.