Multi-source Remote Sensing Fusion For Lithological Information Enhancement

Due to its wide coverage,easy access and highly economic efficiency,remote sensing imagery has been broadly applied to geological exploration,such as structural interpretation,alteration extraction and lithological mapping,especially in arid and semi-arid regions.With continuous development and progress of sensor and computer technology,multispectral remote sensing imagery develops from low-medium resolution multispectral to high spatial resolution,high spectral resolution and high temporal resolution.Lithological discrimination from remote sensing data in geology also develops from human computer interactive interpretation to automated classification.As an important data support for ESA's Earth Observation program for environmental monitoring,Sentinel-2A multispectral imagery has been widely employed in environmental monitoring.However,due to lack of SWIR data containing spectral characteristics of altered minerals,geological applications of Sentinel-2A data are limited.In addition,due to the inconsistent spatial resolution between the Sentinel-2A image bands,some bands lack spatial details.As a result,how to improve the overall resolution of Sentinel-2A imagery through image fusion to make full use of the data is one of the focuses of current application research.In this study,multispectral original Sentinel-2A imagery was comprehensively evaluated in lithological visual interpretation,alteration extraction,and lithological classification in geological exploration,by comparing with common geological remote sensing imagery ASTER and OLI.Then advanced image fusion methods were successfully applied to this data,and the resulting fused images were fully evaluated in rock and mineral extraction.The main works and achievements of this study are as follows:(1)Compared with the commonly used remote sensing geological data ASTER and OLI,multispectral Sentinel-2A imagery has better ability to recognize lithological units in in exposed area of bedrock such as Shibanjing ophiolite complex because of its spatial resolution advantage.In addition,the DEM data plays an important role in lithological classification.(2)The band combination of Sentinel-2A and ASTER partially compensates for the spectral defects of Sentinel-2A in the short-wave infrared range.It can be used for the extraction of serpentine alteration information,and also improves the capability of Sentinel-2A image for lithological discrimination,such as carbonate rock,silicate rock and sulfate rock.(3)the inconsistency of the Sentinel-2A image band spatial resolution was solved by the Multivariable Analysis(MV)image fusion process,which improves the spectral resolution of Sentinel-2A image in the visible near-infrared and short-wave infrared range,and improves the ability of Sentinel-2A to extract iron-bearing lithological unites and minerals.(4)The image fusion between Sentinel-2A and hyperspectral data Hyperion improves the spectral resolution of Sentinel-2A image,and improves the capability of Sentinel-2A data for lithological and mineral information extraction.In this study,multi-level image processing technologies such as image information enhancement algorithms,image classification algorithms and image fusion algorithms were used to mine the potential of Sentinel-2A image to extract geological information,and an effective information extraction method from Sentinel-2A data for geology was developed.