Study On Accuracy Evaluation Of Multi - Scale Remote Sensing Monitoring Of Rocky Desertification In Southeast

Karst rocky desertification is an extreme form of the deterioration of the ecological environment is also the current our country sustainable development in southwest China is facing a very serious problem. Remote sensing technology because of its wide range of observations, rich information, fast, and the characteristic of the dynamic monitoring in time, become today’s monitoring of rocky desertification is the most effective technological means. The development of remote sensing technology to produce the mass of the earth observation data, makes the field of remote sensing monitoring is more and more widely. How to make the application characteristics and advantage of different data sources and different scales, different remote sensing data sources of rocky desertification monitoring research, and scientific and reasonable evaluation about the quality of the monitoring results, has become one of the highlights in agro-scientific research in the current rocky desertification remote sensing. Based on SPOT5simages based and MODIS, TM in three different resolution of remote sensing data, the rocky desertification of remote sensing information extraction test respectively, and evaluate its precision and correlation analysis. Of the paper research content mainly embodied in the following aspects:(1)With karst rocky background combines southeastern Yunnan, select three remote sensing data at different spatial scales, and from the factors spatial resolution, temporal resolution, coverage and price and other data characteristics were analyzed to determine the different remote sensing data scope of the study.(2) For three different resolution of remote sensing data, according to the study of the practical need of and the characteristics of their data, studied the test data preprocessing, respectively. For SPOT5simages based data and TM data mainly has carried on the radiation correction, geometric correction, and band synthetic processing, and projection for MODIS data is mainly for the data and data format conversion processing research. (3) With the normalized difference vegetation index (NDVI) as the basic parameters, by using the principle of binary model based on pixels, calculated the test area of vegetation coverage, respectively, on the basis of combining with the indexes of land use status quo, on the study of karst rocky desertification information extraction using remote sensing technology.(4) In the study error matrix methods were used to evaluate the accuracy and area of karst rocky desertification monitoring results of multi-scale remote sensing properties of geometric precision and accuracy were evaluated and analyzed.