A new RS classification algorithm is proposed where classification takes place in multiple steps. Initially, easy to classify pixels are detected, which are then spatially processed to create additional contextual information to assist classification of the more challenging pixels.
Further information: G. Mountrakis, L. Luo (2011). Enhancing and replacing spectral information with intermediate structural inputs: A case study on impervious surface detection. Remote Sensing of Environment, 115(5):1162-1170.[1.8MB pdf]