Oil Spill Detection Based On The Compact Polarimetric SAR

Oil spill is one of the most important sources of marine pollution.Synthetic aperture radar(SAR)is able to detect marine oil spills due to its all-weather and all-day imaging capability.However,conventional algorithms using single channel SAR usually confuses oil spills with oil spill lookalikes,which makes oil spill detection a challenge task.In recent years,polarimetric SAR have been studied to improve the oil spill detection performance,in particular with respect to distinguishing between oil spills and lookalikes.Compact polarimetric(CP)SAR is one of important trends of polarimetric SAR platforms.In analogy to the full polarimetric SAR,the compact polarimetric SAR contains abundant polarization information of the scattering target.What's more,it also has the characteristics of large imaging range,and the cost of manufacture is much lower as well.Therefore,compact polarimetric SAR is a promising tool for oil spill detection.In this paper,we utilized the compact polarimetry HP-mode data and derived three parameters to distinguish oil spills from lookalikes.The three novel parameters include Compact polarimetric entropy(Hc),compact polarimetric fraction(PFc),and compact polarimetric pedestal height(PHc),which are conducted from the compact polarimetric eigenvalue decomposition.The scattering type of the sea surface is dominated by low entropy scattering,i.e.,Bragg scattering from small rough surface,which can be considered as weakly depolarized.Based on former consideration,the Hc is high,the PFc is low,and the PHc is high,which means the average scattering is in depolarized state.The target is considered as "random" and the polarization status information is lost.Hence the aforementioned parameters can be used to detect oil spills.In this paper,Radarsat-2 and SIR-C / X-SAR data of C-band were used to detect oil spills.The results show that the Hc,the PFc,and the PHc can distinguish between oil spills and lookalikes effectively.The performance of all the parameters based on the compact polarimetric eigenvalue decomposition is excellent,not only can effectively exclude the lookalikes,highlight the oil spill area,enhance the difference between oil spill and sea clutter,but also can effectively maintain the details.At last,this paper synthesizes the parameters of the compact polarization entropy,compact polarization fraction,compact polarization pedestal height,compact polarization degree,and the ellipticity parameters,and proposes a method to detect marine oil spill using random forest.Compared with traditional polarimetric SAR and compact polarimetric SAR oil spill detection methods,the results show that this method has the best detection effect.The oil spill detection method has good generalization ability and relatively accurate detection capability,and has the potential for large-scale detection of marine oil spills.