|
Water resources is the blood of the ecosystem , the Earth 's environment is the most important and the most dynamic factor . Pollution and eutrophication of lakes, rivers, water bodies is increasing due to rapid economic development , the process of industrialization and other human activities , the impact of global freshwater resources are increasingly failure . Compared to conventional methods of water quality monitoring , water quality monitoring method based on remote sensing technology has a wide range of speed , low cost , and ease of the advantages of long-term dynamic monitoring . Poyang Lake is the study area , based on field water quality sampling , measurements of spectroscopic instruments , MODIS images of three synchronized data . The measured spectra showed Poyang Lake presents common characteristics of class II water body , the lower the concentration of chlorophyll a peak is not obvious , suspended solids has the peak . Based the the image characteristics Chavez black pixel selection atmospheric correction of MODIS images , single band after correction , the reflectivity band combination chlorophyll a concentration , the concentration of suspended solids do peason correlation analysis results showed : chlorophyll a reflectivity . high correlation , there have been a large number of bands and band combinations , the the best inversion band to 500m resolution ( B2 B7) ; suspended solids with spatial and temporal characteristics , with only a small number of the band or band combination significant correlation , the best inversion 500m resolution band ( B2 -B7 ) . Using statistical regression analysis based on the best bands , respectively, and BP networks are two ways to establish the remote sensing inversion model , and the two inversion model test measured sample data . The results show that : the data has a the strong radiation resolution of 500m than 250m data inversion effect ; BP network in a sufficient number of samples used to build the nonlinear mapping model , relative to conventional statistical regression modeling showed a stronger nonlinear mapping ability . Finally, between the measured concentration of water quality parameters the pearson analysis , there are very significant correlation found between the suspended solids and transparency and statistical regression method to get the mathematical expression of the suspended solids and transparency , combined with the Carlson trophic state index and modified trophic state index formula has been calculated based on remote sensing technology, TSI index model . Combined with synchronous MODIS images , application model generation Poyang Lake wet period TSI index maps .
|