RS information extraction techniques of Pinus massoniana forests based on knowledge rules

The application of remote sensing(RS) technique in forestry industry has been restricted by the low precision of the information extraction.In order to improve the precision,three key techniques of the information extraction were studied,including selecting training samples,selecting classification bands and training classifier.Then the method and process of information extraction based on knowledge rules were built.This method was applied and validated in Fuyang City,Zhejiang Province,and the information extraction precision was up to 80.5%, which was 20% higher than that of MLC(maximum likelihood classification) method.The results show that this method is effective in the tree species information extraction and there is still research potential.This method can provide technique support for RS application on forest pest studying of Zhejiang Province.