Improved The Algorithm Of Single Image Dehazing Based On The Dark Channel Prior

With the rapid development of community economy nowadays, an urgentdemand for safety is required for the advance in the living standard ofpeople. As the action of safety protection is becoming more and moreimportant, monitor, as the most powerful measures for security work, hasbeen widely applied. But in the adverse weather conditions, such asraining or fogging, especially the problem of the environment is more andmore serious nowadays, fog haze weather becomes a common phenomenon, thereliability of the monitoring system is facing a huge challenge, thefeature information attenuation often appears in the images acquired bymonitoring. And sometimes, the images might be found completely destroyed.In this way, the required information could not be obtained accurately,which appears as a tremendous hidden trouble in security work. Therefore,it is very important to develop the image dehazing technology research.Based on the Dark Channel Prior Dehazing theory, the weakness of thismethod is compensated in this dissertation for unable to be used in somespecial areas, such as offwhite scenery or bright sky. First, we use somemethods to make the channels of three colors close to zero. Then, threechannels each of which lack of one color are defogging restorated. Bycalculating the weighed sum, the defogging images are obtained. Theexperiment result shows that this method produces better results in thebackground of offwhite scenery or bright sky than the existing ones.Furthermore, the noise generated by the original methods under such conditions is reduced effectively, which enhances the visual effectremarkably.