Pan-sharpening based on multi-objective decision for multi-band remote sensing images

Abstract

Pan-sharpening applies details injection to fuse a multispectral (MS) image with its corresponding panchromatic (PAN) image to produce a synthetic image. Theoretically, the synthetic image's spectral resolution should equal that of the MS image and its spatial resolution is the same as that of the PAN image. However, for existing pan-sharpening methods, the trade-off between the spectral and intensity information in the process of details injection is insufficient, resulting in spatial or spectral distortion of the fused image. In this paper we propose a novel pan-sharpening algorithm based on multi-objective decision for multi-band remote sensing images to improve the quality of the fused image. The proposed method focuses on developing a parametric model from a multi-objective perspective to simultaneously maximize the quality of all the pixels in the fused image. We introduce a details injection approach to enhance the edge and texture of the MS image. We design an efficient spectral fidelity fusion model based on the injected details using spectral modulation to pan-sharpen the MS image. We provide an algorithm based on multi-objective decision to solve this model. The main advantage of the proposed method is that it can provide effective spectral modulation to eliminate the adverse effects of details injection. We conduct experiments on simulated and real satellite image datasets to evaluate the proposed method. The results show that our method achieves superior performance to other state-of-the-art methods.