A new method of vegetation classification based on temporal distribution of vegetation indices

Vegetation classification is the basis for various hydrological and ecological studies. Some vegetation classifications using NDVI data have been studied commonly, however, there are some limitations in classification of vegetation using the average NDVI value, especially, the study with spectrum characteristics of the NDVI is relatively rare. Based on this, this paper introduces a new vegetation classification method which combines the phase with the NDVI distribution on the basis of computing the probability. Firstly, the probability of each vegetation type in a period was calculated with two-dimensional random variable conditional probability. Then we synthesized the image at germination, maturity and senescence periods, and the probability was calculated according to the fact that various vegetation types may have different biometric information at different periods. At last, vegetation classification map was finished on the probability value. The Wenquan area over the Qinghai-Tibet plateau (QTP) was taken as an experimental area. The result showed that the overall accuracy and kappa coefficient was 70% and 0.60, respectively. So the vegetation classification method can classify and discriminate vegetation effectively.