Proba-V performance assessment for forest cover mapping over the Atlantic biogeographical region of Europe
Publication from Digital
EARSeL eProceedings 10, 2/2011 , 1/2011
The Belgian Federal Science Policy Office (BELSPO) has initiated a Preparatory evaluation/ validation Programme for the products of the new PROBA-V satellite to be launched in spring 2012. The satellite shall allow daily monitoring of terrestrial vegetation cover through remote sensing, and shall cover the data provision gap between the closure of the SPOT/ VEGETATION Programme and the launch of the SENTINEL-3 mission. The aim of this study is to evaluate the improvements that PROBA-V will bring along for forest monitoring in the Atlantic Biogeographical Region of Europe, and lies within the objectives of the FM@PROBA-V project (http:\\fm-probav.maich.gr). A representative site in Northern Portugal is selected for this reason. VEGETATION, LANDSAT-5, and MODIS data along with the JRC Forest Cover Map are used to train the classifiers, simulate PROBA-V data, apply the classifiers at 30m, 250m, 1/3 of a Km, and 1Km pixels, and validate the results,
while quantifying the accuracies. Maximum Likelihood (ML), Artificial Neural Networks (ANN), and Spectral Vector Machine (SVM) methods are put to test. Best performing classifier shall then be recommended to examine its robustness against sudden disastrous events, like fire, at the same area performing change detection between sequential dates (before and after the event). The performance of the data and classifiers are being demonstrated, and the preliminary results are discussed.