Development of Accurate Detection/Identification Method of Stricken Residential Buildings Using Radar Porlarimetry

点击翻译按钮获取中文摘要

In this research, we propose a useful technique for strickenman-made objects identification/classification, by using the combination of the NNED(Non-Negative Eigenvalue Decomposition) and the OAC (Orientation Angle Compensation).It is found from the results of the PolSAR image analysis that proposed technique is validfor the purpose, and the “remainder” component generated in the NNED procedure is alsoconsidered as a secondary but effective marker for stricken area identification.Furthermore, it is cleared from the image analysis that the combination of the co-polarizedcorrelation coefficient in LR circular polarization basis and the ratio of the correlationcoefficient to that under Non Reflection Symmetry is very useful for detecting snow-coveredand/or oriented man-made objects.