The Application Of HJ-1A Hyperspectral Image In The Environment Dynamic Monitoring In Loess Hilly Region

Our society is in a period of rapid development, and with the growth of economic and the expansion of population, for human the demand of resources has increased dramatically. people's irrational exploitation of resources and damage to the environment cause great changes to land use status, vegetation cover and environment form. Also a series of ecological and environmental problems such as soil erosion, desertification and environmental pollution has become even more prominent. Remote sensing technique can monitor the change of land use and its environment promptly and effectively, then update relevant space geographic information data in time, which can provides effective technical support and decision-making basis for the country's ecological construction. Hyperspectral data with its high spectral resolution and the "combining image with spectrum" feature, can be more precise classification of surface features and improve the accuracy of environmental change detection. In this paper, based on the environmental mitigation satellite's hyperspectral (HJ-1A/HSI) data on Baota in Yan an city, I research the technology method of ecological environment dynamic monitoring, and give a analysis to the ecological environmental dynamic changes in this area.The pretreatment of hyperspectral data, in particular the correction of atmospheric radiation is a prerequisite for post-analysis applications. In this paper, a variety of atmospheric correction methods were compared, and their applicability were analyzed too. I used FLAASH model which inherit from radiative transfer model to give an atmospheric correction for hyperspectral data in the research area. After correction, the spectrum curve of the surface features well remove the influence of atmospheric absorption, and it confirmed the validity of the atmospheric correction. At last, reach a conclusion that in the absence of field synchronous measurement data, use atmospheric radiative transfer model method to do atmospheric correction for hyperspectral data is feasible, also it's the most accurate method.The accuracy of remote sensing dynamic monitoring depends on the accuracy of remote sensing image classification, so the selection of classification method is essential. In this paper, I did experiments of classfication with several typical classification methods with the hyperspectral data of the environmental mitigation satellite's about the research area. And the overall accuracy of the SAM method reach 84.57%,especially the accuracy of woodland, grass and cultivated land are 95.94%,88.71% and 86.36%,which, to a certain extent, meet the request of the accuracy of ecological environment dynamic monitoring in large scope in loess hilly-gully area.Finally, the paper determined the surface feature type of the dynamic monitoring object, and interpreted remote sensing image in the Baota area. Then combined with the dynamic investigate results from ETM image in 2000, we established a dynamic monitoring database for the result, and implemented ecological environment change dynamic monitoring based on the HSI hyperspectral data in Bao Ta area and completed the processes of ecological environment dynamic monitoring based on HJ-1A/HSI hyperspectral data. The methods and result in this paper provides a reference in our new type hyperspectral imager(HSI) data applications in the dynamic monitoring of ecological environment, meanwhile it give a foundation for more application in other fields in the future.