Research On The Methodology Of Synergetic Application Of Aster And Worldview-2Images On Lithologic Information Extraction

Remote sensing technology can provide the rocks, minerals, and geochemical information of earth surface, so there have been various kind of remote sensing data are widely used in litho-logy identification and deposit exploration in recent decades. This paper considered the charact-erristics of the ASTER and worldview-2data on the spatial resolution,spectral resolution and radiometric resolution, analyzed the pros and cons of the two data, then draw the conclusion that because of ASTER data have shortwave infrared (and thermal infrared) bands which are sensit-ive to the rocks and mineral information,so it has a better response to the lithological information, and worldview-2data can only detect visible-near infrared information of earth surface,so it has a relatively weak spectral ability to identify lithologies. However, by worldview-2’s ultra-high spatial resolution, it can detect more of the spatial structure and texture features of the rock units.On the basic of of these analysis, introduce the concept of scaling, then collaborate the adv-antages of the two kinds of data on the spectral scale and spatial scales, explore the techniq-ues and methods of synergetic application of medium and high resolution remote sensing data on lithological information extraction.In the spectral scale, to take the improvement of the spectral detection capabilities of the two datas to the earth surface as principle,using the western section of the western Kunlun orogenic belt-Kurgan Tage region for the study area, stacking the non-overlapping band of ASTER and worldview-2reflection data in spectral range, and apply the BP neural network supervised classification method on ASTER, worldview-2and stacking data respectively to extract the lithologic information of9lithologic units in study area.. Though the comparative analysis of lithologic classification results of stacking data, ASTER and worldview-2data,found that the classification accuracy of stacking data is higher than ASTER classification accuracy on6kinds of lithologic units of study area, and higher than worldview-2data on8kinds of lithologic units of study area.Furthermore,we found that the overall classification accuracy of ASTER data is higher than worldview-2data.. The results show that the synergetic applications of spectrum of the ASTER and worldview-2data can improve the respective accuracy of the classification on lithologic extraction, and it serves as an effective means.In spatial scales, considering the scale effects of the surface processes, synergetic the applic-ation of spatial structure and texture features reflected by multi-scale remote sensing images to abtain the comprehensive description of surface phenomena. Using the western section of the western Kunlun orogenic belt-Hantiereke region as the study area,resampling the panchromatic band with0.48m spatial resolution of worldview-2data to2m,5m,10m,15m and30m respect-ively, then apply the Gray-level Co-occurrence Matrix method to extract the texture features on these image with diffirent scale,and analyze the lithologic information of earth surface revealed by multi-scale texture, taking the texture feature as new band to stacking with multi-spectral bands of ASTER and worldview-2data,performance the supervised classification method on sta-eking data,extracting the4kind of sandstone information of Hantiereke area. The results show that the multi-scale image texture information can reflect the process and pattern of the surface at different scales, and diffirent texture features can improve the multi-spectral data classification accuracy on diffirent degrees, can be used as an important auxiliary information to synergetic application with spectral information to used in lithology information extraction.