Research And Analysis Of Iron Tailings Identification By Hyperspectral-Multispectral Remote Sensing

It is an important means for dynamic monitoring by RS to monitor mine environment and restore management evaluation.There are various types of tailings deposits in our country and emission increased year by year,which cause environment pollution more and more seriously.It is hard for traditional identification technology and extraction accuracy of tailings by RS to meet requirement of precision monitoring in mining area.It is significant for dynamic monitoring mines by RS to use hyper-spectral feature to establish identification window for rapid extraction of iron tailings.This paper studies effect of tailings size and humidity on RS spectral interference zone.Spectral measurement was performed.Spectral parameters were extracted,and spectral of different samples were matched with SAM and SFF after average,wavelet denoising,envelope removal and other pretreatment.Results show that reflectivity increases with decrease of size.Spectral feature at 0.40?m,0.55?m,2.22?m and 2.3?m,which has higher matching score and better fitting effect,are not affected by size and dry states.Spectral measurements with different deposit in Hebei were carried out to identify and analyze its differences.It was found that there are similar spectral curves and absorption feature reflection spectrum in different areas with same deposit and common spectral identification marks with different deposits after preprocessing,feature parameters extraction and matching analysis.Typical spectral absorption peaks at 0.55?m,0.95?m,1.94?m are less affected by associated minerals.Spectral absorption at about 0.55?m,which is less affected by granularity and wet-dry state and accompanying minerals,can be used as identification window for precise extraction through above experiments.Extraction experiment and accuracy analysis with 0.55?m identification window were carried out by using GF-1 in Yinshan mining,in which extraction accuracy of dry tailings,open mines and mountains increased by 22.04%,4.49%,and 19.59%.It verified validity of iron tailings spectral identification window for multi-spectral RS information recognition,which is of certain reference value for actual RS monitoring and precise information extraction of mining area.