论文部分内容阅读
针对遥感图像融合中,不同地物区域对空间与光谱信息要求不同的问题,提出了一种基于显著性分析的自适应遥感图像融合算法。结合多尺度谱残差分析模型,将遥感图像分为纹理、边缘丰富的显著区域与纹理、边缘较少的非显著区域,对显著性不同的区域采用不同融合算法。针对居民区、道路等纹理、边缘信息丰富的显著区域,采用窗均值亮度色调饱和度(IHS)变换,较好地保留了空间细节;针对农田、山地等非显著区域,采用基于小波变换的融合策略保留较多光谱信息。实验结果表明,新算法能使融合结果中的显著区域保留更多空间细节,非显著区域保留更多光谱信息,为今后的遥感图像融合研究提供了一定的理论与应用价值。
Aiming at the problem of different spatial and spectral information requirements in remote sensing image fusion, an adaptive remote sensing image fusion algorithm based on saliency analysis is proposed. Combined with multiscale residual analysis model, remote sensing images are divided into texture, edge-rich salient areas and texture, edge-less non-salient areas and different fusion algorithms for different salient areas. Aiming at the salient areas with rich texture and edge information, such as residential area, road, and so on, the spatial detail is preserved by window-average IHS. For the non-salient areas such as farmland and mountain, the fusion based on wavelet transform The strategy retains more spectral information. Experimental results show that the new algorithm can preserve more spatial detail in the salient region of fusion result and retain more spectral information in non-salient region, which provides some theoretical and applied value for the future research on remote sensing image fusion.