薄松散层覆盖煤矿采空区高光谱遥感特征研究

The coal goafs become incompact and (sub)oxidizing environment after mining,conduce sulfate and carbonate oxidation,coal spontaneous combustion,and gas resolution from surrounding bed,these changes brings goafs physcial and chemical field response.This research takes 2-3 coal field goafs covered by unsolidated layers as experiment area, chooses those typical points at different geology structure and deformation position over the goafs,firstly, to collect the mix hyper-spectrum of these points by ASD FieldSpectra whose spectrum ranges from 350nm to 2500nm, and then to gather the gas rising from the goaf mainly including sulfur dioxide,sulfureted hydrogen,carbon monoxide,methane,also to test the surface tempture.The content of each component of the gas from field will be tested in laboratory but the content of niton is tested at the locale by radio trace device. The next job is to study the characters of the hyper-spectrum from those data mentioned above.The spectrum is the electromagnetic energy emitted by objects at the surface,which includes reflection and radiation ,maybe mixt by different pure spectrum.So there are some relationships between the spectrum and surface objects,the objects is composed of the gas rising from underground goaf,this statement is approved by most scientists through their previous study.But the physical mechanisms and the behavior how the gas influence the electromagnetic spectrum is still discussed. The first goal of this study is to find out the steady portions of the mixt spectrum,the steady portions are response to the gas according to the composition contents,those are indications for goaf existing.The route and arithmetic includes absorbtion position,absorbtion width,absorbtion area etal.calculation aiming to get the spectrum character parameters,the regression analysis between field spectrum parameters and pure gas spectrum parameters is implemented. The next goal is to study the variational law of the steady portion influenced by different gas contents,the variation is concerning at the stength of steady portion.The third goal is to discover the unsteady portion and the rule how it behaves along with the gas contents chaning.The quantity relationship between them is the main interest. Finally,the physical mechanisms leading to steady portion and unsteady portion in the spectrum above goaf should be discussed.Take the characters of goaf spectrum as the base, this study improves the methods of hyper-remote sensing image processing to detect goaf,develops some new ways to extract information from image about goaf. This study and the results may help to detect subground goaf quickly and cheaply,also provide data for understand the mechanism of electromagnetic spectrum mixing.

薄松散层覆盖煤矿采空区因采动效应和自燃效应，将引起地表相应的物理化学场变化，导致地表光谱异常，这些异常的变化规律以及引起光谱异常的原因，目前尚不清晰。本研究拟通过分析煤矿采空区高光谱特征参数，定量表达这些变化规律，分析其变化原因。以华北薄覆盖型煤田2－3个已知采空区为实验区，选择采空区不同地质构造位置、受力变形位置，针对采空区产生并运移到地表的物质和能量，主要是二氧化硫、硫化氢、一氧化碳、甲烷和氡气，测试其含量，并采集对应点的高光谱曲线，构建高光谱特征参数，研究这些参数变化特征与采空区地表气体物质成份及其含量的关系，查明对采矿区探测有指示意义的"稳定谱段"及其变异规律性，研究其定量关系和物理机制，以此为基础，研究基于特征谱形的高光谱遥感识别方法的改进途径。 该研究有助于认识煤矿采空区高光谱变化产生的机理，提高遥感对采空区的探测效果，具有科学和生产意义。

本文利用EO-1卫星HYPERION高光谱影像，进行合适的数据预处理，提取自然水体和塌陷区水体的高光谱曲线，进行等间隔重采样，利用一种高光谱曲线的全波形分维算法，识别采煤塌陷区积水和自然水体，结果表明，该方法能够从光谱角度有效识别两类水体，自然水体的分维数要大于塌陷区水体，通过系统聚类，利用单一的分维数分类精度要高于利用原始光谱曲线的分类精度，可为塌陷区的动态监测提供有效手段。. 由于项目执行时间较短，已投稿论文两篇，录用一篇，成果还未刊登。今后继续深入研究，所刊发成果以该项目资阻为标识。