Image And Video Enhancement Under Insufficient Lighting Conditions

In the real world, people always are jammed by insufficient lighting conditionswhen they capture images and videos. Obtained images and videos’ quality isreduced. This situation seriously affected the people’s lives and travel, securitymonitoring, research and so on. Therefore, research of enhancement algorithm forimage and video obtained under poor lighting conditions has been of concern in thefield of image processing with important theoretical significance and practicalvalue.In daily life, there is three kinds of insufficient lighting conditions includingfog light conditions, night lighting and underwater lighting conditions, which cancover the vast majority of a variety of insufficient lighting conditions in the realworld. Previous degraded image or video enhancement algorithms proposed by mostof researchers are often only accommodate one kind of insufficient lightingconditions and it’s hard to obtain a uniform framework for image and video’senhancement under insufficient lighting conditions. Accordingly, proposing anautomatically enhancement algorithm for images and videos obtained underdifferent kinds of insufficient lighting conditions has important practicalsignificance.Though the research and analysis on the characteristics of the insufficientlighting conditions, we find some similarities between the three. For this reason, wepropose a new enhancement algorithm for image and video obtained under badlighting conditions, which can automatically identify the lighting conditions underwhich the input data obtained. Then, apply different enhancement algorithms fordifferent illumination conditions. The main contents are as follows:Firstly, we put forward an image and video enhancement algorithm for foggyweather conditions. Through improving dark channel prior and introducingbrightness correction method, achieve the purpose of real-time foggy image andvideo enhancement while ensuring the quality of enhanced image and video.Secondly, based on an important observation between the night image andfoggy image and statistical analysis result, we present a real-time night image andvideo enhancement algorithm. Through improving the foggy image enhancementalgorithm by introducing multi-scale median filtering operation, the new algorithmcan enhances the visibility of details and overall brightness of night image and videowhile maintaining the naturalness of the result and avoiding excessive enhancementPhenomenon.Thirdly, for the features of underwater images and videos, we present a underwater image and video enhancement algorithm based on atmosphericscattering physical model. By modifying the overall color of the image and video,enhancing the contrast and adjusting the brightness, we can eliminate the presenceof color shift and haze to achieve the purpose of enhancing underwater images andvideos.