Application Of Dual-polarization SAR Data In Paleochannels Detection

Synthetic aperture radar(SAR) remote sensing data have many advantages whichoptical remote sensing data do not have. It not only has the ability of high resolutionearth observation, and penetrability.The micro topography, especially the space targetposition and morphological characteristics have better discriminant ability.Polarization radar data comprehensive recorded the phase information of objectscattering echo and amplitude information in the form of a scattering matrix. Itprovides more polarization information for target detection,scattering analysis andclassification.By studying the features of backward scattering characteristics andpolarization characteristics,can reflect the terrain roughness, water content etc., forgeological and geological phenomena (concealed), especially the paleochannelsidentification has a unique advantage.Paper preliminary try to apply ALOS PALSAR data.By using ALOS PALSARdual polarization data,making further study of dual polarization SAR data processingand data features, typical features in the study area, polarization and scatteringproperties in jilin in the western region.On the basis of above,comparing the differentpolarization SAR data effect and precision of the method.Finally extracting paleochannel information of research area, show the cause of the water system changesmade in this paper.Main work and results reflected in the following respects:1.The imaging principle of SAR image is different from optics image, it will beseriously affected by the coherent spot.To improve the processing precision, the studyadopt multi look processing and filter for noise reduction.After dealing with multilook processing, the radiation resolution of images was improved.Lee, Frost,enhanced Lee were used respectively to three kinds of filtering algorithm of imagefiltering.By comparing three methods of standard deviation,ENL indicators,we foundthese three filtering methods could effectively suppress the speckle noise,andenhanced Lee is a method which is the best for the paper.2.Extracting polarization parameters, combined with the field test, to study inseveral typical features of backscatter features, as well as in HH, HV, span image donein-depth research on different polarization characteristics.For the realization of thefollow-up work of the SAR image of high precision classification provides credible classification rules.3.Based on typical features in-depth analysis and summarization of thepolarization scattering properties, paper have used the decision tree classificationmethod of SAR image classification.compared with other different classificationmethod (K-Means), the classification of different types of images (Landsat8) as aresult, we found that classification result based on the features of the polarizationcharacteristics of dual polarization radar data is better than optical remote sensingimage.4.On the basis of classification, combining with the field test work, access to alarge number of historical literature data,paper have extracted the research of thepaleochannel information.By analysising reasons of nenjiang river and songhua riversystem change in the studied area, the conclusion shows that the crust uplift in theeastern part of songliao basin is within the scope of the study area paleochannelformation of the real driving force.The L-band polarization radar data can effectivelydetect the paleochannel in this area.