Spectral Characteristics Of Typical Saline-alkali Patch And Remote Sensing Identiifcation

Xinjiang is one of the concentrated distribution areas of salt-affected soils in China. Many researchers insaline soil formation evolution and classification, and probably the remote sensing monitoring and forecastof research has accumulated a large number of research results. But there are many " Saline-alkali patch "phenomenon in oasis farmland, seriously affected the crop yield, reducing the effective arable land. Inthis study, the research material was farmland soil in Typical Saline-alkali patch in Manasi of Xinjiang.After sampling and determining of soil spectra in field, we analyzed the relationship between soil salinityand soil spectral curve characteristics. Using continuum-removed and first derivate methods, the sensitivebands which reflected soil sainlization level were selected. Hyperspectral quantitative prediction model wasconstructed. Based on TM images, By soil salinization remote sensing modeling method based on featurespace, the salinity in the studied area identified by using remote sensing monitoring index model (SDI), inthe study area, saline soil identification made saline spot in the study area map, Main conclusions are listedbelow,(1) Typical Saline-alkali patch spectral curve center spot overall spot patterns and trends peripheralsalt is consistent, at a wavelength of1450nm, a significant absorption peak near1940nm.(2) Generally the Saline-alkali patch spectral curves changed gently, spectral characteristics weresimilar in morphology and were parallel substantially. which exhibited a trend that the higher the saltcontent was, the higher the reflectivity.(3) Through correlation analysis among normalized spectrum, the first derivative spectra and soilsalinity, we found that the sensitive bands which reflected soil salinization level involved650-700nm, Thespectrum prediction models of normalized and the first derivative were established, i.e.,y=61.113x691+0.457. Through the inspection found a model of order differential treatment is better.(4) by using the TM remote sensing image, the comprehensive analysis of the normalized differencevegetation index (NDVI), salt index (SI) the relationship between them, on this basis to establish NDVI-SIfeature space concept, and build the saline soil spot remote sensing monitoring model index (SDI), Theresults showed that the soil salinity and SDI correlation is higher, the R2=0.8596. Passes the inspectionaccuracy, and also show good distinguish SDI to soil salinity spot in the study area distribution range. SDIcan reflect, saline soil surface salt combination and its variation with clear physical meaning, and the indexis simple and easy to obtain, saline spot information extraction and dynamic monitoring in arid areas in thefuture research is of great significance...