VEGETATION CLASSIFICATION BY MULTI-SCALEHIERARCHICAL SEGMENTATION ON GF-2 REMOTE SENSING IMAGE

Further classifying the vegetation into subclasses is significant for some applications such as ecological protection and vegetation mapping. A method based on multi-scale segmentation is proposed in this paper to separate the GF-2 remote sensing image into objects on different hierarchies. On each layer, high-resolution images are separated by using spectral, shape, texture and other image features. Experiment on the data of area along the Yangtze River in the Dianjun District of Yichang City shows the efficiency of the method. When the shape factor fixed to 0.1 and three layers used with fuzzy membership classifier and k-nearest neighbour classifier, respectively, the kappa coefficient of the final vegetation classification reaches 0.97.