基于极化特征SERD的SAR溢油检测

Oil Spill Detection Based on Polarimetric Feature SERD

与单极化SAR(Synthetic Aperture Radar)相比,全极化SAR图像中不仅包含散射目标的几何特征和后向散射特征,还包含散射目标的极化特征。因此,基于极化特征的SAR图像分类能够更全面地描述海面目标的物理特性。单次反射特征值相对差异度(Single Bounce Eigenvalue Relative Difference,SERD)能够比较单次散射机制的相对大小,并且可以反映散射表面的粗糙度情况。而海面油膜的存在抑制了海面的短重力波和毛细波,改变了海表面的粗糙度。基于此,本文将SERD应用到海面溢油检测中。利用两景Radarsat-2全极化SAR数据对比分析了SERD与极化散射熵的溢油检测效果,实验发现:(1)SERD能够较好地区分溢油与海水。(2)对原油而言,SERD的油水对比度与极化散射熵的油水对比度在数值上差异较小;对生物油膜而言,SERD的油水对比度在数值上远小于极化散射熵。利用这一特性,SERD在区分生物油膜与原油方面更具优势。

Compared with single polarimetric SAR（Synthetic Aperture Radar）,full-polarimetric SAR images contain not only geometrical characteristics and backward scattering characteristics of the scattering targets,but also the polarization features of the scattering targets.Classifications of SAR images based on polarimetric characteristics can better reflect the physical property of sea surface targets.SERD（Single Bounce Eigenvalue Relative Difference）is related to roughness of sea surface and is able to measure the proportion of single scattering mechanism.Moreover,slicks on the sea surface dampen the gravity-capillary waves,changing roughness of the surface.Based on the above theory,we apply SERDto oil spill detection.The classification results are compared with polarimetric entropy（H）.Two Radarsat-2full polarimetric SAR images are analyzed.The experiments show that：（1）SERDcan effectively distinguish oil spills from sea background;（2）for crude oil,SERDhas a similar slick-to-water contrast with polarimetric entropy;for biogenic film,the slick-to-water contrast of SERD much less than polarimetric entropy.Hence,we can distinguish plant oil and crude oil by taking the advantages of this property.