We propose an unmixing framework for enhancing endmember fraction maps using a combination of spectral and visible images. The new method, data fusion through spatial information-aided learning (DFuSIAL), is based on a learning process for the fusion ...