Research On Lithological Classification Using SAR Image And Its Application On Geology

In recent years,the SAR techniques are developing fast, deepening itsapplication on geology. SAR image has a good performance on detection ofgeological structure and rock mass. SAR image can be used to extract informationwhich is not include in optical images and this special ability could be more helpfulfor recognition of hiding geological structure, mineral characteristic, undergroundwater and so on. The study area, the Huangshan area, is located in the Xinjiang VygurAutonomous Region, and physically the area has a typical geomorphic feature ofgobi.In this paper, four different kinds of polarization SAR data of RADARSAT-2will be used.In the phase of pre-process of data, the NEST software is used to filter, calibrateand re-project the SAR data.During the filter process, the result gained by using fourtype filter methods are analyzed, such as Lee、Gamma map、Frost、Refines Lee, andit can be found that the LEE filter has the best performance among all the filtertechniques.In the phase of application of data, analyzing the SAR image characteristicpresented by four different kinds of polarization modes, it can be found that the crosspolarization data has a variety of advantages over other ones to present the surfacefeatures. During the work of interpretation, analysis is developed by using theRADARSAT-2VH polarization data, and in this paper I established the typicalfeatures of different of rock mass and the different geological strata. For the main, theinterpretation work should be based on the tone and texture structure. The resultsproved that the rock and mineral has light tone, while the stratum has dark tone, as thetopography has no important impacts on them. Different rock mass’ form and texturefeatures have great difference, so we can distinguish rock mass of different kinds oflithology focusing on form and texture.In the phase of field work and labratory data processing, the surface roughness parameter is measured by a common and simple surface-roughness-surveyor, meanwhile some mineral samples are also collected. During the labratory data processing, the dielectric coefficients of rock mineral samples collected during fieldwork are measured.In the phase of retrieval of surface roughness from SAR images and validation ofresults, during this part, the surface roughness is retrieved from the cross polarizationimage (VH polarization image subtract HH polarization image), based on which the work of rock and mineral classification is developed, combination with the fieldwork, we verified the results of classification being present the distribution of basic rock, marble, granite and sedimentary rock. Besides, based on the analysis using spatial-texturemethod, we get the results of classification of granite, basicrock and ultrabasicrock distribution. At last, analyzing the classification results of four different kinds of polarization image, we get the results of distribution of granite using image enhancement method, and the results of distribution of diorite, granite and something of quaternary system. According to the fieldwork, the results are validated by in site data. It has been proved that the results are good.