Research on the influence of noise properties on sea oil detection with PolSAR remote sensing imagery.

Noise properties of polarimetric SAR (PolSAR) imagery have received more and more attentions in recent years, especially in marine oil spill surveillance. The property of noise is an important indicator to assessment the quality of the backscattered signals. In this paper, three Radarsat-2 PolSAR imageries of six selected areas including oil and ocean surface are investigated. Analysis results of dominant noise type over those target areas suggest that the behaviors of phase difference distributions are closely related to the noise properties. The co-polarized complex correlation coefficient ρco is more like an indicator of the dominant noise type instead of various sea surfaces. Moreover, A linear correlation between the μ, DoP, H and the corresponding ρco, indicating the performances of those parameters are closely related to the noise property. Besides, the cross-polarized complex correlation coefficient has the potential to distinguish oil-covered sea surface from its surroundings under high wind condition. In addition, the noise properties probably contribute to non-Bragg scattering over oil slick for low quality PolSAR imagery, yet the features of the oil slick for high quality PolSAR imagery are still dominated by Bragg scattering. [ABSTRACT FROM AUTHOR]