Performance Comparison of Dehazing Algorithms on different platforms

Images captured may suffer from attenuation due to the transmission medium and scattering of the air-light. When the weather is foggy the scattering is more, so the attenuation is also to a larger extent resulting in degraded image. Most of the automated and computer vision based systems (like image classification and retrieval, vehicle navigation aids, remote sensing etc.,) which depend on the image captured outdoor may not work effectively with the degraded images. Several researches have been done to enhance the foggy images captured outdoor and indoor. Amongst them Dark Channel Prior with guided filtering is a traditional method, while Color Attenuation Prior is a linear model developed based on the relation between depth of the scene, brightness, concentration of the fog and saturation of the image. But an algorithm will be effective and efficient if it can be implemented with available resources and consumes less processing time. We compare in this paper the performances of two dehazing algorithms using different platforms like Windows and Unix and different tools like Matlab and OpenCv python. The comparison shows that OpenCv python implementation is much faster and cost effective than Matlab tool since Matlab is a proprietory tool whereas OpenCv python is an open source.