Remote sensing image fusion (or pan-sharpening) aims at generating highresolution multi-spectral (MS) image from inputs of a high spatial resolutionsingle band panchromatic (PAN) image and a low spatial resolutionmulti-spectral image. In this paper, a deep convolutional neural network withtwo-stream inputs respectively for PAN and MS images is proposed for remotesensing image pan-sharpening. Firstly the network extracts features from PANand MS images, then it fuses them to form compact feature maps that canrepresent both spatial and spectral information of PAN and MS images,simultaneously. Finally, the desired high spatial resolution MS image isrecovered from the fused features using an encoding-decoding scheme.Experiments on Quickbird satellite images demonstrate that the proposed methodcan fuse the PAN and MS image effectively.
展开▼