基于可见-近红外和热红外光谱联合分析的煤和矸石分类方法研究

发布时间：2018-04-28 20:23

本文选题：煤 + 矸石　；参考：《光谱学与光谱分析》2017年02期

【摘要】：煤与矸石是矿山采煤过程中主要固体堆放物,对其进行遥感动态监测是矿山环境保护的重要需求。由于煤与部分矸石存在"异物同谱"现象,在使用传统的可见-近红外遥感分类时,往往将部分矸石划分为煤,导致遥感分类精度降低。首先对铁法矿区的12个煤样本和115个矸石样本进行可见-近红外光谱测试,发现绝大部分矸石样品的光谱与煤差异很大,二者易于区分,但有部分矸石与煤样本存在"异物同谱"现象。为进一步对矸石与煤区分,测试了混分样本的热红外光谱,发现二者存在明显的光谱差异,利用热红外光谱特征可以将其区分开来。在此基础上,提出了基于可见-近红外和热红外光谱联合分析的煤与矸石区分方法。该方法首先对所有样本进行可见-近红外光谱测试,利用Mao模型进行第一步分类识别;其次对煤与矸石混分的样品进行热红外光谱测试,利用光谱吸收比率SAR作为判别指标进行第二步分类,两步的分类结果为最终分类结果。该方法在铁法、兖州、神东和木里矿区的验证结果表明,其具有很高的分类准确率,效果远好于单独基于可见-近红外光谱特征的分类方法。研究结果表明,利用多种光谱联合分析的方法可以解决单波段存在的"异物同谱"现象,对于地物遥感分类具有重要的借鉴意义。
[Abstract]:Coal and gangue are the main solid stacks in the mining process, and it is an important requirement of mine environmental protection to monitor them by remote sensing. Due to the phenomenon of "foreign body isospectral" between coal and some gangue, some gangue is often divided into coal when using traditional visible-near infrared remote sensing classification, which results in the reduction of classification accuracy. First of all, 12 coal samples and 115 gangue samples from Tiefa mining area were tested by visible near infrared spectroscopy. It was found that the spectrum of most gangue samples was very different from that of coal, and the two were easy to distinguish. However, some gangue and coal samples have the phenomenon of "the same spectrum of foreign bodies". In order to further distinguish gangue from coal, the thermal infrared spectra of mixed samples were tested, and it was found that there were obvious spectral differences between them, which could be distinguished by using the characteristics of thermal infrared spectrum. On this basis, a method for distinguishing coal from gangue is proposed based on the combined analysis of visible-near-infrared and thermal-infrared spectra. In this method, all samples are tested by visible and near infrared spectroscopy, and the first step is classified by Mao model, and the samples mixed with coal and gangue are tested by thermal infrared spectroscopy. The spectral absorption ratio (SAR) is used as the discriminant index for the second step classification, and the result of the two-step classification is the final classification result. The results of verification in Tiefa, Yanzhou, Shendong and Muli mining areas show that this method has a high classification accuracy and is much better than the classification method based on the characteristics of visible and near infrared spectra alone. The results show that the phenomenon of "foreign body isospectral" in single band can be solved by using the method of combined spectral analysis, and it is of great significance for the classification of ground objects by remote sensing.
【作者单位】：东北大学资源与土木工程学院;中国矿业大学物联网(感知矿山)研究中心;
【基金】：国家自然科学基金项目(41440032,41371437)资助
【分类号】：O657.33;TD849

【相似文献】