Vegetation Factors Extraction And The Mothod Of Rapid Determination Of Soil And Water Erosion In Red Soil Erosion Area

Nowadays soil and water erosion becomes the global concerned environmental problem. It is well known to all that vegetation is the active factor of preventing erosion. And increasing vegetation coverage is the best and the most effective way that restraint erosion deterioration. Vegetation coverage is one of the measure indexes of ecological environment, the important parameters of depiciting ecosystem, and the major factor of influencing soil and water erosion. The obtainment and measurement of vegetation coveris the key and hot topic of ecology ang global change field study.It has short cycle, and it is fast, accurate, timely to use the methods of remote sensing image of vegetation coverage for dynamic monitoring. In remote sensing measuring method of vegetation coverage, vegetation index method is widely used. But in southern red soil watershed region, this method still lack of quantitative research. So the key and difficult point of vegetation coverage remote sensing extracting is to choose the most suitable for the use of studied area, the most sensitive to vegetation coverage and the most insensitive to background factor's vegetation index. On the basis of vegetation index, using the improved dimidiate pixel model calculated the vegetation coverage, and applied to the soil and water erosion measurement extraction. It appears very necessary to establish soil erosion rapid extraction measurement model in neural network. That appears very necessary for southern area frequent soil erosion monitoring, and regional sustainable development.This paper takes multiband remote sensing image ALOS 2009 as the basic image data source, combines remote sensing technology and traditional method, remote sensing calculation and investigation, qualitative research and quantitative calculation, uses interdiscipline, more data integration, varieties of methods. And emphasizes the selection of appropriate vegetation index of ChangTingXian ZhuXi watershed in Fujian province, calculates vegetation coverage using the dimidiate pixel model. Then establishes soil and water erosion model utilized neural network method in ZhuXi watershed. The purpose is to promote rapid extraction model applied to southern red soil erosion areas.The results showed that the most appropriate vegetation index for the ZhuXi watershed is MSAVI-Modified Simple Vegetation Index, and the method of estiminating vegetation coverage by remote sensing method is feasible, the precision can reach a certain requirements. In metlab, the present situation of soil and water erosion simulation and predict research in the watershed performs well in the BP neural network. Keywords:Vegetation Index, the dimidiate pixel model, Artificial Neural Networks, the Zhuxi watershed...