Improved segmentation of a series of remote sensing images by using a fusion MRF model

Classifying segments and detection of changes in terrestrial areas are important and time-consuming efforts for remote-sensing image repositories. Some country areas are scanned frequently (e.g. year-by-year) to spot relevant changes, and several repositories contain multi-temporal image samples for the same area in very different quality and details. We propose a Multi-Layer Markovian adaptive fusion on Luv color images and similarity measure for the segmentation and detection of changes in a series of remote sensing images. We aim the problem of detecting details in rarely scanned remote sensing areas, where trajectory analysis or direct comparison is not applicable. Our method applies unsupervised or partly supervised clustering based on a cross-image featuring, followed by multilayer MRF segmentation in the mixed dimensionality. On the base of the fused segmentation, the clusters of the single layers are trained by clusters of the mixed results. The improvement of this (partly) unsupervised method has been validated on remotely sensed image series.