Cartographier par télédétection l’occupation du sol et ses changements

Abstract

This article presents a method of mapping by remote sensing of land cover and its changes based on the case of Southern Senegambia. The retrospective approach adopted is based on the comparison between the so-called "reference" spectral signatures resulting from ground truths and those of the classes resulting from the unsupervised multispectral classification of images taken at earlier dates. The result is a cartography of the main land cover categories at each of the dates under study and their evolution over time. The main changes in the Senegalese forest landscape were thus highlighted: a significant decline in land afforestation in the north of the region over the last decade and a regeneration of the mangrove following the improvement in rainfall conditions.

Keywords

• Land cover change mapping
• Landsat image
• multispectral classification
• land use
• Southern Senegambia