Evaluating the effectiveness of fusing remote sensing images with significantly different spatial resolutions for thematic map production

Publication date: Available online 18 September 2018Source: Physics and Chemistry of the Earth, Parts A/B/CAuthor(s): Shaohui Chen, Weimin Wang, Hong LiangAbstractIt is very challenging to fuse remote sensing images with significant spatial resolution difference. In this paper, such kind of images are fusing with one posteriori probability model for the purpose of creating thematic maps. The outlined solution consists of parsing the low spatial resolution multispectral images into intensity-hue- saturation triples with spatial and spectral properties separated; the contribution of the old intensity and the panchromatic image is factorized into posteriori probabilities using their wavelet planes under the aid of support vector machine; the wavelet planes are probabilistically weighted to foresee one new pan-sharpened intensity, from which the fused images are produced by inverse IHS. The outlined method along with two representative methods are evaluated with two sets of imagery covering contrasting heterogeneous landscapes for different purposes of road extraction and wheat area map.

Source: Physics and Chemistry of the Earth, Parts ABC - September 20, 2018 Category: Science Source Type: research