无人机高光谱岩性和矿物识别技术研究  

Study on lithologic character and mineral identification technology based on UAV hyperspectral image

作　　者：刘洪成[1] 叶发旺[1] 韩晓青[1] 谭宏婕 鲁纳川 张川[1] LIU Hongcheng;YE Fawang;HAN Xiaoqing;TAN Hongjie;LU Nachuan;ZHANG Chuan(National Key Laboratory of Remote Sensing Information and Image Analysis Technology Beijing Research Institute of Uranium Geology,Beijing 100029,China;China University of Geosciences(Beijing),Beijing 100083,China)

机构地区：[1]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 [2]中国地质大学(北京),北京100083

出　　处：《世界核地质科学》2024年第3期560-572,共13页World Nuclear Geoscience

基　　金：中核集团集中研发项目“第四代铀矿勘查关键技术研究与示范”(第一阶段)项目:“新型关键勘查技术装备研制”(编号:遥SD03-03)资助。

摘　　要：无人机高光谱遥感技术以其高空间分辨率和高光谱分辨率的技术优势,在大比例尺区域地质填图领域可以发挥重要作用。利用无人机搭载全谱段高光谱成像仪(HD_SAVI)获取0.2 m高空间分辨率影像数据,采用经验线性法对所获取数据进行光谱重建,重建后无人机高光谱影像光谱曲线与地面实测光谱曲线谱形基本一致,典型特征吸收峰位置偏差均在1个像元以内。采集岩性样本构建岩性知识库,利用深度学习算法基于无人机高光谱影像数据开展岩性识别;岩性识别结果与大比例尺地质图对比,岩性空间分布基本一致,地质体边界准确且较清晰。参考典型矿物光谱库进行矿物识别。与同一地区CASI/SASI航空高光谱数据地质填图结果进行对比分析,结果表明:两者提取的矿物在空间分布上具有较高的一致性,但无人机低空高光谱影像空间分辨率较高,识别出的矿物种类更加丰富。无人机高光谱遥感技术可以满足无人区、交通不便地区地质勘查,以及重点区或矿区外围等地区大比例尺地质填图的需要。The unmanned aerial vehicle(UAV)hyperspectral remote sensing technology has achieved rapid development in geological exploration due to its versatility and flexibility.In this paper,The UAV was equipped with a full-spectrum hyperspectral imager(HD_SAVI)to capture hyperspectral image data,The image reached a spatial resolution up to 0.2 meters and was radiometrically corrected using the empirical linear method with the standard reflectance material deployed in the survey area.The corrected hyperspectral image was highly agree with ground-measured spectra and the deviations of typical feature absorption wavelength within one pixel.Lithological sample spectra were collected to set the knowledge base for the rock type identification by deep learning algorithms.Mineral identification was performed using a standard mineral spectral library,and structural identification was carried out using edge detection operators.These was followed by a comparative analysis with results from CASI/SASI airborne hyperspectral geological mapping in the same region.The results indicated that the minerals extracted from both hyperspectral datasets show a high level of consistency in spatial distribution.The low-altitude hyperspectral imagery captured by the UAV system has a higher spatial resolution,allowing for the identification of a larger scale mapping of rock types and minerals.The comparison of lithology identification results with large-scaled geological maps shows that the spatial distribution of lithology is basically consistent,and the boundaries of pluton was relatively clearer.Therefore,unmanned aerial vehicle hyperspectral remote sensing technology can meet the needs of geological prospection on areas of depopulation and inaccessiable transportation,and the large-scaled geological mapping in key areas or their pepherical areas.

关 键 词：全谱段高光谱成像仪 经验线性法 岩性样本 深度学习 大比例尺地质填图 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置] P619.14[自动化与计算机技术—控制科学与工程]