Low-pass filtering based polarimetric dehazing method for dense haze removal.

Polarimetric dehazing method is very promising in enhancing the quality of images captured in the scattering media. However, it is found that the dehazing results calculated by hazy images are very sensitive to the noise, which may cause the method unstable or even invalid. To overcome this drawback and enhance the capability and stability of the polarimetric dehazing method, digital image processing algorithms or bias parameters need to be added into the method, however, they will make the algorithm complex and time consuming. In this paper, using low pass filter to suppress the noise of the hazy images, a novel polarimetric dehazing method is proposed to enhance the visibility of hazy images, especially for dense haze removal. Experimental results demonstrate that this method is totally automatic and very effective in dense haze processing. This method may have great potential usage in many applications, such as optical surveillance, underwater imaging, and bio-tissue imaging, etc.