Study On The Remote Sensing Technique Of Monitoring Agricultural Drought In North China Based On HJ-1 Hyperspectral Data

The technology of remote sensing is one of the means to monitoring drought. Compared with traditional technology of monitoring technology, remote sensing technology has the advantages of macroscopic characteristics, timeliness and economical efficiency. The technology of remote sensing can provide spatial, spectral and time information of large area, multi-scale, multi-temporal land surface, and can objectively reflect the characteristics of the integrated information of land surface. The use of remote sensing technology to monitoring wide range of regional drought is of great significance on relevant departments of the decision-making and guidance of agricultural production. Hyperspectral remote sensing can reflect both spatial and spectral information, and help to improve the accuracy of information extraction, having great potential of monitoring agricultural drought. In this paper, HJ-1 hyperspectral data is used to monitor exemplarily the drought in the region of Chifeng, Inner Mongolia, based on the approach of hyperspectral remote sensing.Atmospheric correction and spectral reconstruction of hyperspectral remote sensing applications is the key technology and necessary prerequisite. This paper summarizes three methods of atmospheric correction, and atmospheric radiative transfer model is the most accurate method. The hyperspectral data of Hyperion is be used to contrast and demonstrate the atmospheric radiative transfer models, both MODTRAN4.0 and 6S, in field of optical remote sensing. Then the software of FLAASH, basing on MODTRAN4.0, is used to implement atmospheric correction and spectral reconstruction of the data of study area.The spectral resolution of hyperspectral data facilitates band selection on the model of retrieval. Using reflectivity after atmospheric correction data, the retrieval of vegetation index, fraction of vegetation, leaf area index, chlorophyll content index and other drought parameters are fulfilled, and vegetation supply water index is retrieved combining with the temperature based on HJ-1 infrared data. Then monitoring of drought is fulfilled by comparing the results of retrieval in study area of different time.The types of surface features in study area are established by the results of retrieval of parameters and spectral analysis. Then "red edge" of vegetation spectral is calculated by program. Then spectral angle mapping based on end-member spectrums by way of spectral clustering is fulfilled in study area, further demonstrated the effects of drought. As aspects of the above, the processes of monitoring drought based on HJ-1 hyperspectral data are established, laying a good foundation for further study and application.