Detection Of Pine Forest Area Changes In Regions Damaged By Bursaphelenchus Xylophilus In Zhejiang Province

Pine forest area changes in regions damaged by Bursaphelenchus Xylophilus in Zhejiang province from1994 are estimated using RS techniques. Total 12 scenes of TM images of 3 scenes at different places and 4 times and some field forest inventory data have been used in this research. The results are much consistent with the facts.A new method called linear transform is proposed. 2 or more bands that selected for classification are sorted, both from small to larger or from large to small, and the lightness value is taken as the dependent variable and the order number as the independent variable to build a linear regression model. Then the slope and the intercept are used as classifying features. This new method has been employed in separating water boy and non-water body built-up and non-built-up, conifer and non-conifer, and good results have been gained. Especially in the first two cases the precisions are much higher than the maximum likelihood estimation method. The advantage of this method is not only reducing the data dimensions but also making convenience for analyzing at a 2-dimension feature space.A high-dimension non-parameter lookup table method is proposed. The first step is to select valuable bands, the second is to build single land-use type high-dimension histograms and to smother them, the third is to transform the histograms into probability values, and the last is to transform the probability histogram into the lookup table.Some good combinations of bands are proposed. Such as TM2, 3, 4, 5 for water-body and non-water-body, TM3,4,7 for built-up and non-built-up, and TM4,5 for conifer and non-conifer.According to data of 12 scenes of the spectrum of pine and fir curve analysis concluded that only rely on the spectral information of TM data is not sufficient to distinguish between these two species. So the paper taken advantage of Forest resources survey data of Zhejiang provience as auxiliary data to estimate the pine forest area.The pine/conifer forest area ratio is estimated to be multiplied by the remote sensing of coniferous forest area, then will receive the of the forest area estimates by remote sensing of coniferous forest distribution. According to the remote sensing estimates of the Bursaphelenchus Xylophilus disease in Zhejiang Province for 13 years against the region between the forest area decreased by 26.1%.Changes in confier forest area were not only caused by Bursaphelenchus Xylophilus disease but also by other reasons , such as in recent years Zhejiang Province promoted the " confier forest to broad-leaved forest " activities will also lead to reduction in confier forest area. The real number of area reduced by Bursaphelenchus Xylophilus disease needs deeper investigations and studies.