Fast Algorithms for Foggy Image Enhancement Based on Convolution

摘要

When drivers moved in foggy weather, traffic accident often happened because of poor vision. Many research works have been done to this subject by scientists. One of these is the Retinex theory, according to which, only the information reflected the objects' own characteristics are preserved, and some uncertain factors, such as the intensity of the light and the non-uniformity of the irradiation, are thrown off. Based on this hypothesis, many algorithms for foggy image enhancement have been developed. But a large computation works need to be done when these algorithms were used and in many situations, such as vision enhancement for foggy traffic, real-time performance is required. To solve this problem, a new algorithm FSSR for this topic, which is based on SSR algorithm, was proposed. The algorithm FSSR keeps the good enhancement results, and computation task is reduced evidently.