Underwater Dark Channel Prior Method Of Image Enhance

Marine science and technology has become the focus of world with the economicand technological development. For the purpose of underwater exploration andconstruction, wide range and accurate underwater environmental data are needed. Inorder to obtain these data, a series of ocean technologies were under researched. Andthe underwater imaging method is researched in this article. Underwater image isimportant for scientific research and technology. But due to the absorptive andscattering nature of seawater, underwater images are essentially characterized by theirpoor visibility and color distortion. In this paper we propose a new image processalgorithm which easily removes degradation effects in underwater vision.Dark channel prior method can effectively improve the quality of the degradedfoggy image remove the effect of fog. And this paper analyze the relationship anddifference of underwater water and foggy image and present the idea of remove thisalgorithm to underwater image. The research contents of this dissertation are asfollowings:The underwater image imaging process and the underwater image existing problemis analyzed; Analyses the advantages and limitations of the underwater imageprocessing method.Introduces the dark dark channel prior algorithm, and puts forward this method inthe underwater. Experiments prove that the original dark channel method cannot bedirectly transplanted into underwater. Through analysis and contrast calculation, theeffects of underwater dark channel algorithm is mainly color distortion andunderwater light uneven illumination.Because the dark channel prior algorithm based on the value of RGB color channels, and underwater image have color degradation problem. In view of theunderwater color distortion problem, this paper proposes using the dynamic thresholdwhite balance algorithm to restore the underwater images.This method is based on theanalysis of the images YCbCr color space to determine the reference points, choosingthe reference white point threshold is dynamically.Through the analysis of the imagesof the Cb Cr coordinate space, get a close to white contains the reference white pointarea (near region). Base this method find out the reference white point, and thenthrough the von-kvies model to adjust the picture period to achieve the white balance.And the we use the underwater24color standard color plate and two underwaterobject for processing images, thus proves that the automatic white balance algorithmbased on dynamic threshold is effective.To solve the problem of uneven underwater image lighting. We analyze theunderwater optical image model. Through the actual measured value and the biggestgrid value of image itself, using underwater image imaging formula to estimate thebackground light, and use it in underwater dark color image enhance. At the sametime compare the improve algorithm and original algorithm, prove that the effect ofunderwater dark channel method is better than the original algorithm in underwaterimage..Pool experience: In order to objective of the image, the contrast, the informationentropy and average gradient of image processing are used. And then using the poolexperiment, in the artificial environment, extraction under the clean water, normalquality water, and turbidity water quality to take underwater images. And useunderwater dark channel image enhancement algorithm processing the images whichobtained under three kinds of water quality, on the subjective evaluation ofunderwater dark color image processing algorithms processed images is better thanoriginal image obviously. And the contrast, information entropy, average gradient ofprocessed images are better than the original images. Through these data we candemonstrate that the effectiveness of underwater dark color image enhance algorithm.Marine experiment: first introduced the inhomogeneous illumination underwater image detection system, using it to take the images in the extraction of natural waters.And then use the underwater dark channel image enhance algorithm to process theseimages, by observed, the conclusion underwater dark color image enhance algorithmis more excellent to correct the color of the image and remove the scattering noise, theprocessed image with high resolution, high contrast, high resolution, etc. At the sametime, through the image contrast, the information entropy, average gradientcalculation is further improve. It mean that the advantage of the dark color underwaterimage enhance algorithms.