The Early Identification Of Remote Sensing About Bursaphelenchus Xylophilus Based On Process Model

The Bursaphelenchus xylophilus is belong to a disaster which is aroused by invasion of exotic forest pest. It has very serious impact on forest health, ecological security and Social economic in our country. The Pinus massoniana which are destroyed greatly by the Bursaphelenchus xylophilus in recent years, is one of the main tree species in Fujian province. It is of great meaning to monitoring the Bursaphelenchus xylophilus disease.This paper choose Fujian province as research area and Pinus massoniana detroyed by Bursaphelenchus xylophilus as research object to do the studies using early remote sensing recognition with some spectral measurement and moisture content determination in different suffer period with SPOT5 remote sensing image in January 2009. According to the typical vegetation spectral differences in the study area, we classfied the vegetation using Decision Tree based on Knowledge Rules in the study area. And then extract the information of pinus massoniana. By analyzing the mechanism of the Bursaphelenchus xylophilu, the article proposed that the leaf water content is the effectively index of the early detection and discussed the relationship between Spectrum information (Vegetation Index) and water content in leaves, providing basis for the early detection of Bursaphelenchus xylophilus. Finally, according to the actual situation, we choose the relatively common Spectral Angle of classification in high spectral remote sensing with a measured masson pine spectral data of Bursaphelenchus xylophilus disease of different period as the reference sample library and do the sample recognition with radiation corrected SPOT5 multi-spectrum remote sensing image in different suffer period of masson pine. The study found that because of the influence of the climate and other factors, we need a large number of sample data as reference if we regard leaf water content factor as disease recognition factor. So we need a further research on the early classification recognition of disease. Spectral Angle classification can be more effectively recognize different period of victims samples, but classification threshold setting and identification accuracy still remained to be discussed further.