干旱区生态环境知识资源中心

A desertification monitoring model was constructed and applied in a region nearby Gaotai Country, China, where undergoing a typical desertification process. First, the surface albedo and vegetation index were retrieved. And by statistical regression analysis, the quantitative relationship between albedo and NDVI of different desertification land was determined, and found they have good correlation with the coefficient 0.7707. Based on this, the model of desertification monitoring was constructed; then, using decision tree to classify the desertified land and evaluate the quantitative relationship of them by natural breaks method; finally, Quick Bird image was used to test the classifying results and the overall accuracy of the model reaches 84.071%. The study shows that the indexes of the model can reflect the desertification land surface cover, the water-heat combination and their changes. It can make full use of multi-dimensional remote sensing information and can achieve the automatic identification of desertified land. (C) 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of Conference ESIAT2011 Organization Committee.