In this paper, we propose a novel remote sensing fusion approach based on guided image filtering. The fused images can well preserve the spectral features of the original multispectral (MS) images, meanwhile, enhance the spatial details information. Four quality assessment indexes are also introduced to evaluate the fusion effect when compared with other fusion methods. Experiments carried out on Gaofen-2, QuickBird, WorldView-2 and Landsat-8 images. And the results show an excellent performance of the proposed method.
This paper proposes a Markov random field (MRF) model with adaptive selection multiresolution (MRF-AM) for texture image segmentation. By considering the wavelet decomposition and the morphological wavelet decomposition, MRFAM adaptively selects the multiresolution representation as features from the wavelet and morphological wavelet stepby- step. Then, the MRF is employed to model the features of adaptive multiresolution. The segmentation results are finally obtained by maximizing a posterior probability of the MRF. Experiments demonstrate that our method can improve the segmentation accuracy compared with the deterministic multi-resolution method.