Hyperspectral and multispectral image fusion based on constrained CNMF unmixing

Hyperspectral (HS) remote sensing image with finer spectral information has great advantages in feature identification and classification. However, the spatial resolution of HS image is usually low due to physical limitation and data transfer requirement. In this paper, the low-spatial-resolution HS image is fused with the high-spatial-resolution multispectral (MS) image of the same observation scene to improve its spatial resolution. A novel spectral unmixing based HS and MS image fusion approach (VSC-CNMF based fusion approach) is proposed, in which CNMF with minimum endmember simplex volume and abundance sparsity constraints is employed for coupled unmixing of HS and MS images. The fused image is built by the product of endmember signature matrix derived from HS image unmixing and fractional abundance matrix derived from MS image unmixing. Simulative experiments illustrate that compared with the CNMF based fusion technique, the newly proposed VSC-CNMF based fusion algorithm is capable of producing fused images closer to the reference image with better spectral fidelity.