An approach for land cover change detection using low spatial resolution data

Abstract

The paper investigates an approach for detecting land cover changes on low spatial resolution data. On a reference year, local curves describing the temporal reflectance in the visible and near Infrared related to each land cover type are computed from low spatial resolution data. Temporal profiles of NDVI are calculated and analysed to extract relevant features characterizing them for each land cover type. The same process is then applied regularly on the same site for the next years. The observation in the same features representation space allows detection of potential local change.