Using The GAF Transform and MODIS Time-Series to Perform Landcover Classification and Change Detection

The Gramien Angular Field (GMF) transform encodes time-series into images so that image based classification approaches can be exploited to perform time-series classification. Settlement expansion is one of the most pervasive forms of landcover change in sub-saharan Africa. In this paper, we show that the GAF transform outperforms the conventional techniques we compared against when it is employed in conjunction with Moderate Resolution Imaging Spectroradiometer (MODIS) time-series to detect settlement expansion in the Gauteng province of South Africa.