Research Of Typical Material's Hyperspectral Information Extracting Method In Shuichang Iron Mining Area Located In Eastern Of Hebei

There are a large number of mines, and the types are miscellaneous. The eco-environment in mining area and circumambience has deteriorated during the mining operations. In order to command the mine environmental status, new techniques of investigating mine environment quickly is searching after badly. We choose Shuichang iron mine in eastern of Hebei as the investigation area, take use of Hyperion data, test and analyze the typical materials' spectrums and the chemical elements, dive into the identifying and distilling methods of typical objects' hyperspectral information, investigate the environmental status of our research area and circumambience. We have gained some results and cognitions like below through this research.(1) All of the spectral reflectance of ore, mullock, wall rock, gangue, and surface soil in the mining area has an absorb limit in the special absorbing wave bands of iron, which illuminates that there are quite quantitative iron element in the materials here. Magnetite is the principle iron-bearing mineral in this mining area, and hematite takes the second place. There is a power negative correlation between reflectance and TFe content of ore in the band region from850nm to 900nm. The relation coefficient is 0.9606, and the regression equation is y= 0.3051x^-0.3357.(2) The spectral curves of the vegetation in the area have the effects of "downward of the green peak, up-shifts of the red valley, blue shifts of the red edge" universally, which is the typical aberrance feature of heavy metals. Iron element content in vegetation here all exceeds 600mg/kg, which is on the high side at large compared with the same vegetation in non-iron areas. From industry field to periphery, iron element content in the same vegetation manifests a trend of decrease. There is also a quantization relationship between red edge and the iron element content in the vegetation in the mining area. Thereinto, there is a quadratic polynomial negative correlation between red edge and iron element content of pine, whose relation coefficient is 0.9993, and the regression equation is y= -0.0589x² + 43.038x.(3) After atmospheric correction of Hyperion data, we can subtly identify and distill the iron-bearing mineral information and vegetation's biochemical effect information by use of hyperspectral hourglass processing. Endmembers' apparent reflectance extracting from the calibration spots is basically similar to the normal spectrum and the road testing spectrum, and the spectral parameters are inosculate reciprocally.(4) Taking magnetite as research object, we have founded remote sensing estimate quantization index base on the spectral reflectance and the TFe content of ore grade. We also have investigated the distribution status of ore in the mining area through hyperspectral remote sensing. The investigating results indicate the following status. The ore whose TFe content is more than 20% is mainly distribute in the mid-west of the Beishan strip stope, and the bilateral TFe content of ore grade is mainly between 10%～20% of the stope, which takes on a trend of leanizement gradually. The investigate results by remote sensing is accord with the ore's distribution.(5) Taking red edge location of vegetation as index, we investigate the environmental status in the mining area by hyperspectral remote sensing. The results have approved the following status. The vegetation growing in the industry field and the neighborhood appear the effect of blue shifts of red edge, which illustrate that the environment of the industry field and the neighborhood is polluted universally by iron element.The keystone of this research is the identifying and extracting method of typical objects' hyperspectral environmental information, and the analyzing results provide academic and datum gist for upgrading the veracity and practicability of remote sensing investigation of mine environmental pollution.