Vegetation Classification Model Based on High-resolution Satellite Imagery

摘要

Based on a SPOT-5 image, this study built knowledge pool of vegetation spectral information, adopted classification algorithm of decision tree, proposed a vegetation classification model based on their spectral information and classified the vegetation of Nanjing. The results showed that the overall accuracy was 86.95% and Kappa coefficient was 0.8287. Then the classification model was validated by using an IKONOS image of Yuhuatai region and was improved through combining the textural information. The classification overall accuracy was increased to 92.70% and Kappa coefficient was increased to 0.8648.