Extracting granite pegmatite information based on GF-2 images and the random forest algorithm
DU Xiaochuan1,2(), LOU Debo2(), XU Lingang1, FAN Yinglin1,2, ZHANG Lin1, LI Wanyue1,2
1. School of Earth Sciences and Resources, China University of Geosciences (Beijing), Beijing 100083, China 2. MNR Key Laboratory of Metallogeny and Mineral Assessment, Institute of Mineral Resources, Chinese Academy of Geological Sciences, Beijing 100037, China
Granite pegmatites serve as a significant carrier and prospecting marker of granite pegmatite-type lithium deposits. The southeastern Zhaka area in Tianjun County, Qinghai Province demonstrates considerable prospecting potential for lithium deposits. Nevertheless, its high altitudes and deep cross-cutting characteristics pose challenges in surface surveys. Hence, this study extracted the granite pegmatite information within the study area from remote sensing images using the random forest algorithm. With high-spatial-resolution GF-2 remote sensing images as the primary data source, it extracted the spectral, texture, exponential, topographic, and edge features from various ground objects within the study area. These features, together with the newly introduced contrast limited adaptive histogram equalization (CLAHE) features, constituted 25 feature variables, forming a feature subset. Then, feature variables in the subset were evaluated for their feature importance, and their importance scores were used for feature selection, determining the optimal feature combination for extracting granite pegmatite information. Ultimately, 16 feature variables were chosen for random forest classification, with the accuracy of the classification results assessed. The study indicates that: ①The CLAHE feature variables emphasize the tonal variations among ground objects, thereby enhancing the classification accuracy, with the overall accuracy increased by 2.7 percentage points and the Kappa coefficient increased by 0.035; ②The classification results for granite pegmatites based on GF-2 images and the random forest algorithm exhibited overall accuracy of 93.1%, with a Kappa coefficient of 0.902, user accuracy of 94.24%, and producer accuracy of 98.00%, confirming the effectiveness of the method used in this study. Moreover, this study provides reliable data for future research in the study area.
杜晓川, 娄德波, 徐林刚, 范莹琳, 张琳, 李婉悦. 基于GF-2影像和随机森林算法的花岗伟晶岩提取[J]. 自然资源遥感, 2023, 35(4): 53-60.
DU Xiaochuan, LOU Debo, XU Lingang, FAN Yinglin, ZHANG Lin, LI Wanyue. Extracting granite pegmatite information based on GF-2 images and the random forest algorithm. Remote Sensing for Natural Resources, 2023, 35(4): 53-60.
Wang D H. Study on critical mineral resources:Signficance of research,determiniation of types,attributes of resources,progress of prospecting,problems of utilization,and direction of exploitation[J]. Acta Geologica Sinica, 2019, 93(6):1189-1209.
Wang X Q, Liu H L, Wang W, et al. Geochemical abundance and spatial distribution of lithium in China:Implications for potential prospects[J]. Acta Geoscientica Sinica, 2020, 41(6):797-806.
Yao F J, Xu X W, Yang J M, et al. A technology for identifying Li-Be pegmatite using ASTER remote sensing data in granite of Gobi shallow-covered area:A case study of recognition and prediction of Li-Be pegmatite in Jingerquan,Xinjiang[J]. Mineral Deposits, 2020, 39(4):686-696.
Dai J J, Wang D H, Dai H Z, et al. Geological mapping and ore-prospecting study using remote sensing technology in Jiajika area of Western Sichuan Province[J]. Geology in China, 2017, 44(2):389-398.
Dai J J, Wang D H, Dai H Z, et al. Reflectance spectral characteristics of rocks and minerals in Jiajika lithium deposits in West Sichuan[J]. Rock and Mineral Analysis, 2018, 37(5):507-517.
[7]
Cardoso-Fernandes J, Teodoro A C, Lima A. Remote sensing data in lithium (Li) exploration:A new approach for the detection of Li-bearing pegmatites[J]. International Journal of Applied Earth Observation and Geoinformation, 2019, 76:10-25.
doi: 10.1016/j.jag.2018.11.001
Wang H Y. Research on extracting pegmatite dike information of remote sensing image based on depth semantic segmentation[D]. Beijing: China University of Geosciences(Beijing), 2021.
[9]
Verikas A, Gelzinis A, Bacauskiene M. Mining data with random forests:A survey and results of new tests[J]. Pattern Recognition, 2011, 44(2):330-349.
doi: 10.1016/j.patcog.2010.08.011
Wang S Y, Zhang Y W, Yu Z H. Classification of Honghe wetland remote sensing image based on random forests[J]. Geomatics & Spatial Information Technology, 2014, 37(4):83-85,93.
[11]
Dronova I. Object-based image analysis in wetland research:A review[J]. Remote Sensing, 2015, 7(5):6380-6413.
doi: 10.3390/rs70506380
Chen H Y, Meng L K, Zhou Y. Ice snow and cloud detection in remote sensing images based on random forest[J]. Journal of Geomatics, 2022, 47(2):105-110.
[13]
Feng Q, Liu J, Gong J. UAV remote sensing for urban vegetation mapping using random forest and texture analysis[J]. Remote Sensing, 2015, 7(1):1074-1094.
doi: 10.3390/rs70101074
[14]
Dell'Acqua F, Gamba P. Texture-based characterization of urban environments on satellite SAR images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2003, 41(1):153-159.
doi: 10.1109/TGRS.2002.807754
[15]
Stasolla M, Gamba P. Spatial indexes for the extraction of formal and informal human settlements from high-resolution SAR images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2008, 1(2):98-106.
doi: 10.1109/JSTARS.4609443
[16]
Feng Q, Gong J, Liu J, et al. Flood mapping based on multiple endmember spectral mixture analysis and random forest classifier:The case of Yuyao,China[J]. Remote Sensing, 2018, 7(9):12539-12562.
doi: 10.3390/rs70912539
[17]
Ali A, Pour A. Lithological mapping and hydrothermal alteration using Landsat 8 data:A case study in ariab mining district,red sea hills,Sudan[J]. International Journal of Sciences Basic and Applied Sciences, 2014, 3(3):199-208.
Zhang J H, Yao Y B, Cao N. Prediction of whether precipitation based on decision tree[J]. Journal of Geomatics, 2017, 42(5):107-109.
[22]
Lerman P M. Fitting segmented regression models by grid search[J]. Journal of the Royal Statistical Society:Series C(Applied Statistics), 1980, 29(1):77-84.
Chen H X. Population estimation at the building level based on multi-source data and random forest algorithm[D]. Shanghai: East China Normal University, 2021.
[24]
黄东山. 特征选择及半监督分类方法研究[D]. 武汉: 华中科技大学, 2011.
Huang D S. Research on feature selection and semi-supervised classification[D]. Wuhan: Huazhong University of Science and Technology, 2011.