Fog-degraded Image Restoration Using Local Mean Value And Standard Deviation

In the information age, the demand for information is increasing rapidly. Digital image as the most common information carrier in the social activities is the important way of accessing to information. With the constant development of the science and technology, the digital image processing techniques have been made rapid development and have been applied into all aspects of people’s life and work. Owing to the deterioration of the natural environment, the fog becomes one of the common weather phenomenons, leading to the digital image degradation. Therefore, the haze removal becomes the hot subject for research. In many cases, the complexity of the weather condition makes the haze removal technology research more difficult. At present, in order to solve this problem, scholars at home and abroad put forward many effective methods which have some application limitations.This paper proposed an algorithm of haze removal restoration based on local mean value and standard deviation to restore fog-degraded images. The main contributions of the research work presented in this paper are as follows:(1) This paper concluded that the difference of the local mean and standard deviation reflected the change of the images scene depth through analyzing the characteristics of the local mean and standard deviation of fog-degraded images. The atmospheric scattering model showed that the depth of fog-degraded images scene is consistent with the change of atmospheric scattering light. Hence, this paper estimated the atmospheric scattering light in accordance with the difference of the local mean and standard deviation to remove the haze of image.(2) Based on the above work, this paper improved the method of atmospheric scattering light estimation and adjusted formula to have a good effect on the haze removal.(3) This paper put forward the adaptive haze removal algorithm for the smoothness in different parts of fog-degraded images.(4) In this paper, the brightness of the repaired image color could be strengthened to advance the visual effect.This paper testified the effectiveness and feasibility of algorithm by the objective and subjective estimation methods under the VC+6.0. The results showed that algorithm of defogging could effectively and quickly remove the fog on the image.