机构地区:[1]中国地质科学院矿产资源研究所自然资源部成矿作用与资源评价重点实验室,北京100037 [2]自然资源部深地科学与探测技术实验室,中国地质科学院,北京100094 [3]核工业北京地质研究院,遥感信息与图像分析技术国家级重点实验室,北京100029
出 处:《光谱学与光谱分析》2025年第1期38-44,共7页Spectroscopy and Spectral Analysis
基 金:国家自然科学基金面上项目(42172332);国家重点研发计划项目(2022YFC2905001);中央级公益性科研院所基本科研业务费专项基金项目(KK2102,KK2017);中国地质调查项目(DD20230054,DD20230033);第四代铀矿勘查关键技术研究与示范(第一阶段)项目(遥SD03-03)资助。
摘 要:西藏邦铺矿床是冈底斯成矿带上重要的斑岩-矽卡岩型矿床,矿区东段矽卡岩矿体研究程度较低。高光谱成像技术可以快速获取岩芯样品“图谱合一”的高光谱数据,选取邦铺矿床东段矽卡岩典型岩芯为样品,应用国产核工业北京地质研究院自主研发岩芯扫描仪与国外SPECIM岩芯成像仪,开展短波红外(SWIR)高光谱成像测试及对比分析,揭示典型矿物的短波红外光谱特征,并探讨先进的高光谱成像技术应用于矽卡岩勘探的潜力与优势。对同一批岩芯样品分别进行了国内外仪器的高光谱成像测试,利用最小噪声分离变换以及纯净像元指数进行端元提取,而后结合光谱信息散度法选取不同矿物的训练集和验证集样本,采用光谱角填图法进行蚀变矿物填图。共识别出方解石、绿帘石、绿泥石、角闪石、石英、白云母六种矿物,两种仪器矿物填图的结果大体一致,并反映了不同蚀变矿物之间的组合与分布。SPECIM成像仪对两个钻孔样品的平均填图精度为91.5%,Kappa系数为0.87;核工业高光谱成像仪对两个钻孔的填图精度为75%左右,Kappa系数为0.66。两种仪器获取的高光谱数据信噪比均较高,但是由于核工业仪器的数据空间分辨率相对更低,导致图像中存在更多的混合像元,进而影响了填图精度,其空间分辨率的进一步提高是下一步研发的重点。高光谱成像数据表明ZK0010钻孔绿泥石Fe-OH吸收峰波长变化范围为2252.3~2260 nm,ZK0011钻孔绿泥石Fe-OH吸收峰变化范围为2251.7~2258.5 nm,由上至下均向短波方向移动,表明了绿泥石远离矿化中心而相对富Mg,其高值指示了矿体中心位置。高光谱成像技术相较于传统的点测法可以提供更加丰富的空间信息和光谱信息,为指针矿物的光谱特征参数提取与变化规律分析提供了技术支撑。利用高光谱成像提供的大量光谱数据进行绿泥石Fe-OH峰位统计分析,进而可Bangpu deposit in Tibet is an important porphyry-skarn deposit in the Gangdise metallogenic belt,and the research degree of the skarn ore body in the eastern part of its mining area is relatively low.Core hyperspectral imaging technology can quickly obtain the hyperspectral data of core samples.In this paper,the typical core samples of skarn in the eastern part of the Bangpu deposit are selected,and the short-wave infrared(SWIR)hyperspectral imaging test and comparative analysis are carried out by using the core scanner independently developed by the Beijing Research Institute of Uranium Geology and the foreign SPECIM core imager.At the same time,the SWIR spectral characteristics of typical minerals are revealed,and the potential and advantages of advanced hyperspectral imaging technology in skarn exploration are discussed.We carried out hyperspectral imaging tests on the same batch of core samples with domestic and foreign instruments.We extracted the end members using the minimum noise separation transform and pure pixel index.By consensus,this study identified six minerals:calcite,epidote,chlorite,hornblende,quartz,and muscovite.The mineral mapping results by the two instruments were generally consistent,reflecting the combination and distribution of different altered minerals.The mapping accuracy of SPECIM for ZK0010 and ZK0011 is 91.5%,and the Kappa coefficient is 0.87.The mapping accuracy of the samples by the Beijing Research Institute is about 75%,and the Kappa coefficient is 0.66.The signal-to-noise ratio of hyperspectral data obtained by the two instruments is high.However,due to domestic instruments'relatively low spatial resolution,there are more mixed pixels in the image,which affects mapping accuracy.Further improving its spatial resolution is the key point of the next research and development.The hyperspectral imaging data show that the wavelength of the Fe-OH absorption peak of chlorite in the ZK0010 drill hole ranges from 2252.3~2260 nm,and that of chlorite in the ZK0011 drill hole ranges from 22
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...