Based on the Decision Tree Classification of Remote Sensing Image Classification Method Application

Remote sensing images there are synonyms spectrum and the spectrum of foreign body phenomenon, Traditional classification methods easily lead to the misclassification of surface features and leak points, On the basis of this paper, a decision tree to generate a priori knowledge of the statistical relationship with TM1 to 7 multi-band imaging in the ENVI software support for the data source of the decision tree classification of land cover status quo in Zhengzhou City. Comparative to Remote sensing images of the same area three images at the different time, Studies have shown that the decision tree method is simple, intuitive, and overall classification accuracy of 90.65%.