Comparison of two vegetation classification techniques in China based on NOAA/AVHRR data and climate-vegetation indices of the Holdridge life zone

We have developed a new multi-source data set for integrated analysis of vegetation classification at a continental scale, and applied it in China. Two kinds of supervised classification methods, artificial neural network (NN) and maximum likelihood classification (MLC) algorithms were employed to classify the data set in order to ascertain which method is better for this new data set. Classification results were validated with the same test samples and field samples based on GPS. The accuracy of the classification by NN was better than by MLC.