Wan Yu-Qing, Fan Yu-Hai, Jin Mou-Shun
Geological Exploration Institute of Aerial Photogrammetry and Remote Sensing Bureau, Xi'an, 710199, People's Republic of China.
School of Earth Science and Land and Resources, Chang'an University, Xi'an, 710045, People's Republic of China.
Sci Rep. 2021 Jan 11;11(1):440. doi: 10.1038/s41598-020-79864-0.
A gold-silver-lead-zinc polymetallic ore was selected in Huaniushan, Gansu Province as the study area. Hyperspectral aerial images as the primary information source, ground spectrum tests, and sampling analysis were used as auxiliary techniques. They were combined with large-scale mineral and geological maps and other high-resolution satellite remote sensing images. Hyperspectral remote sensing classification identification and quantitative analysis methods were used to study the main mineral resources and rock mass occurrence. Finally, deposit distribution information was extracted and validated. The results showed that the effective classification methods by hyperspectral images were spectral angle mapping, minimum noise fraction transform, and mixed tuned matched filtering. Based on the ground survey, combined with sampling analysis, the accuracy of classification was 80%. The recognition rate of the main ore body-the iron-manganese cap lead-zinc oxide ore-was as high as 81%. This research showed that hyperspectral remote sensing in this mining area has excellent demonstration effects and is worth completing and supplementing original mineral and geological maps. The targets are important areas for detailed follow-up on mineral resource exploration.
选取甘肃省花牛山的一种金银铅锌多金属矿作为研究区域。以高光谱航空影像作为主要信息源，地面光谱测试和采样分析作为辅助技术。将它们与大比例尺矿产和地质图以及其他高分辨率卫星遥感影像相结合。运用高光谱遥感分类识别和定量分析方法研究主要矿产资源和岩体赋存情况。最后提取并验证了矿床分布信息。结果表明，高光谱影像的有效分类方法有光谱角制图、最小噪声分离变换和混合调谐匹配滤波。基于地面调查，结合采样分析，分类准确率为80%。主要矿体——铁锰帽铅锌氧化矿的识别率高达81%。本研究表明，该矿区的高光谱遥感具有良好的示范效果，值得对原始矿产和地质图进行完善和补充。这些目标区域是矿产资源勘查详细跟进的重要区域。
