As the incidence of oil spills increases, the detection and measurement of oil pollution in the marine environment are receiving augmented attention. Remote sensing is an increasingly important tool for the effective direction of oil spill countermeasures. The most available physical quantities in optical remote sensing domain are the intensity and spectral information obtained by visible or infrared sensors. However, besides the intensity and wavelength, polarization is another primary physical quantity associated with an optical field. While the spectral information tells us about materials, polarization information tells us about surface feature, shape, shading and roughness, and has the potential to enhance many applications in optical remote sensing. During the course of reflecting light-wave, water-surface spilled oil will cause polarimetric characteristic which is related to the nature of itself. Thus, detection of the polarization information for polluted water by spilled oil has become a new remote sensing monitoring method. In this paper, four kinds of oils, they are gasoline, diesel oil, motorcycle oil and soybean oil, were regarded as the experimental samples for polluted water, and the multi-angle spectral-polarimetric instrument was used to obtain the multi-angle near infrared spectralpolarimetric characteristic data of different oil-spilled water specimens. Then, the change rule between polarimetric characteristic with different affecting factors, such as viewing zenith angle, incidence zenith angle of the light source, relative azimuth angle as well as waveband of the detector were discussed, so as to provide a scientific basis for the research on polarization remote sensing for polluted water by spilled oil.
Camouflaged targets detection in complex background is a challenging problem. Spectral-polarimetric imaging can offers spectral information and polarization information from the objects in the scene. Fusion of the spectral and polarization information in the images will result in better camouflaged target identification and recognition. In this paper a novel spectral-polarimetric image fusion algorithm based on Shearlet transform is proposed. Firstly, every polarimetric image in each wave band is decomposed into images of low frequency components and high frequency components by Shearlet transform. Then, the fused low frequency approximate coefficients are obtained with weighted average method, and the fused high frequency coefficients are obtained with area-based feature selection method, so features and details from different spectral-polarimetric images are fused successfully. After that, the kernel fuzzy c-means clustering algorithm is used for camouflaged object separation from its background. Experimental results have shown that better identification performance was achieved.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.