Study On Automatic Extraction Of Remote Sense Information Of Karst Rock-desertification Area

The research on the distribution and cause of Guizhou desertification and environmental comprehensive management of desertification region is essentially mature,but until now there has not been relevant research on how to use Remote sensing image analysis to extract desertification information automatically and quantification analysis based on measuring the characteristics of desertification spectral reflectance with variable backgrounds and grades. This paper aims at exploring a method which uses Remote sensing image to extract object and location information and indicates the difference of the characteristics of different desertification spectral reflectance, providing a theoretical basis for desertification management due to better use of remote sensing technology to accept desertification information. The content of this article is as follows:(1) Research on spectral feature of different rank karst rock-desertification under background terrain feature;(2) Design on automatic extraction model of Remote sense information of Karst Rock-desertification area based on artificial neural networks;(3) To evaluate the accuracy of extraction information about Karst Rock-desertification area.The article propose a new automatic classification model of Karst Rock-desertification based on neural networks combined with spectral features in the study of Yachi demonstrate area. The authors continue to classify by neural network classification and spectrum reflecting properties analysis. Finally, carry out field investigation to improve and validate the classification precision. The results show this method eliminates the disturbance and influence of unwanted objects effectively, so as to improve the classification precision.