Study about alpine cold meadow degradation in the source region of the Yangtze River and Yellow River

摘要

The degradation of alpine cold meadow in the source region of the Yangtze River and Yellow River has been causing very serious ecological and social economic problems. To study degradation of alpine cold meadow in the headwater region from 1994 to 2006, Remote Sensing and the machine learning approach was used. Firstly, a classification system, considering the dynamic evolution process of alpine meadow, was offered. Secondly, the support vector machines approach was introduced into classification of satellite remote sensing images. The experiment results show that, the introduced classification system can express dynamic process of alpine cold meadow degradation efficiently; SVM has obtained better precision than minimum distance and maximum likelihood methods for TM images interpretation. Furthermore, the research indicates that degradation of the alpine cold meadow was most marked from 1994 to 2006, and the area of BS&BR land has expanded rapidly.