Research On Extraction Of Basalt In Jining Shallow Covered Area Using Multi-Source Data

Basalt is a "probe" and "window" for exploring the deep parts of the Earth.Hence,it is significant to master basalt distributions in the studies on tectonics,dynamics of deep lithosphere and geological prospecting,etc.Due to wide sediment coverage of Quaternary,the lithological characteristics and boundaries of Jining basalt are not very clear,making lithological identification a major problem.Recently,geophysical,geochemical and remote sensing techniques have been widely applied to lithological identification in covered areas.However,optical remote sensing data are the information of the Earth surface,and geophysical and geochemical exploration data are discontinuous sampling and small scales.As a result,it is difficult to accurately map lithology based on individual type of these data in covered areas.In order to improve the accuracy of lithological identification in the covered areas,the shallow covered area at Xinghe-Ebping in Jining District,Wulanchabu City,Inner Mongolia Autonomous Region was used as the experimental area for the accurate boundary of the basalt in the covered area in this study.In addition,how to use remote sensing,geochemical,geological and other multi-source data to automatically classify and identify the basalt in the covered area was discussed thoroughly and comprehensively.The main work and results of this study were as follows:(1)Through analyzing lithological visual interpretation,lithological information enhancement and machine learning classification using various multi-spectral remote sensing image data,the optimal data combination and suitable algorithm for lithological identification in the covered area from synergies of multi-source data were found.The results proved that the ability of multi-source data to classify lithology was significantly better than single data’s;and Supporting Vector Machine was the optimal algorithm when multisource data was used to identify lithology in coverage area.(2)The interpolation analysis was performed on the geochemical data of the stream sediments to find the correlation between geochemical elements and basalt in the covered area.It was found that the most representative element combinations of basalt distribution in the covered area are: Cd,Co,Cr,Cu,Mo,Ni,Zn and Pb.(3)Combined the geochemical data and remote sensing data which can effectively reflect the distribution of basalt in the study area,SVM was used to identify basalt.The results showed that the combination of geochemical data and remote sensing data was effective to identify lithology in the covered area,from which high-precision basalt spatial distribution in the covered area can be obtained.In conclusion,the framework and technology of using multi-source geoscientific data proposed in this study can give full play to the advantages of different geoscience data,improve the accuracy and efficiency of lithology mapping in the covered areas,and will provide a technical reference for lithological identification in the covered areas,with important academic significance.