Implementation And Research Of Adaptive Histogram Equalization And Image Enhancement

Image enhancement is a hotpot in image processing research.It improves the visual effect,such as increasing contrast,adjusting illumination,improving sharpness,enriching details,correcting color cast.This technology has wide applied value in monitoring,aerial photography,live,post-production,biomedicine,industry.Through years of development,many image enhancement algorithms have been proposed.But there are some problems exists,such as low universality,overexposure,detail lost.The main propose of this paper is to research the shortcomings of existing algorithms and analyze the character of different images,then implement a adaptive algorithm based on histogram equalization and white balance.The main work of this paper is as follows:(1)Researching the general situation of histogram equalization and white balance algorithms: Summarize the general situation and significance of histogram equalization and white balance.Then analyze and discuss the shortcomings of existing algorithms.(2)Proposed the bi-histogram equalization based on probability weighted uniform distribution model.First,split the histogram into two sub-histograms.Second,get the dynamic segment point by calculating the weighted probability density.Then redistribution the dynamic range of histogram.Finally,post-process the new histogram and process histogram equalization.(3)Proposed a clip histogram equalization using entropy-contrast mixed model.Firstly,increase the tiny histogram value in the whole histogram.Then calculate the entropy and contrast of image by traverse the clip value.Third,achieve the best clip value by balancing two parameters.Lastly,process histogram equalization.Both of the algorithms are high adaptivity.(4)An algorithm of white balance based on achromaticity constancy is proposed to enhance the images in color cast.The algorithm start with calculating the color cast of light source.Then achieve the achromatic weight of every pixel according to the color cast of light source.Finally,correct the color cast by using achromatic weight.The experiment result shows that this algorithm is better than existing white balance algorithm in the filed of protecting the white or black areas.It also apply to many extreme situations such as undersea and sandstorm.(5)Finally,algorithms was simplified and implement on the FPGA platform using Verilog language(1080p,60fps).Experiments have shown that simplified algorithms meet the needs of real-time processing.The effect of output video is still good and universal.