【摘要】 利用高光谱图谱结合特征开展矿山污染直接识别研究 .首先详细分析了德兴铜矿矿山污染 (废矿、废水以及植被 )地物的光谱特征 ,总结出可利用于直接识别和提取这些污染物的特征光谱 ,从而利用矿区航天Hyperion高光谱数据并以矿物识别谱系技术为主有效地识别出矿区的污染类型及其分布 .对于以黄铁矿等含铁矿物为主的围岩或贫矿矿石的氧化污染利用 70 0nm、10 0 0nm以及 2 2 0 0nm附近的特征吸收分别识别出含Fe3 + 矿物及其Fe2 + 和Fe3 + 混合矿物 ,并进一步根据光谱特征识别出赤铁矿和针铁矿 ;根据矿区水体在 6 0 0nm附近吸收特征的差异相对区分出酸性水、碱性水和中性水 ;根据植被在 6 85nm附近的最大吸收深度相对地划分植被污染程度 .最后建议建立矿山污染地物光谱数据库 .该研究为利用高光谱的技术优势快速且有效地直接识别与提取出污染源的种类、类型并分析其潜在的污染趋势提供了新的思路 ,为矿山污染监测、治理规划和复垦提供了新技术和知识支撑 .更多还原

【Abstract】 The process of contamination identification at Dexing copper mine based on the spectral feature of pollutions such as mine offal, waste water and vegetation and so on are investigated using spectral identification tree technique for Hyperion data. The spectra of various surface materials at mine are analyzed at first. And then the different contaminations, the Fe-bearing minerals including Fe³⁺ and mixture of Fe²⁺ and Fe³⁺ based on the spectral absorption feature of 700 nm, 1 000 nm and 2 200 nm, the pollution water and their relative pH based on the spectral feature of 600 nm, the vegetation contamination caused by mine offal and pollution water based on the maximum absorption of spectral depth between 580 nm-750 nm, are identified and extracted using Hyperion data. The spectral database of mining pollution is proposed. A good idea of identifying mining pollution quickly and directly is put forward using hyperspectral imaging technique. The project can be very practical in terms of technical support for inspecting and surveying, managing and planning, remedial action of mine environment.更多还原