Research And Realization Of Improved Histogram Equalization Algorithm

Histogram equalization has become a research hotspot in the filed of image enhancement due to its low computational complexity,simple and effective methods.Although current improved algorithms of histogram equalization solves the defect of traditional histogram equalization to some extent,they still have some problems in practical application.For example,the algorithms of brightness maintenance have limited application range,while the algorithms of brightness variable mostly have fixed parameters,so they are not good for some images.The main purpose of this paper is to research and implement the improved histogram equalization algorithm that can be used for video real-time enhancement in various scenarios.The main work of this paper is as fllows:(1)A bi-histogram equalization algorithm based on maximum entropy model is proposed.Firstly,the data segmentation point of histogram is determined by otsu method.Then the histogram is preprocessed.After that,the segmentation point of the dynamic range with the maximum entropy after histogram equalization is searched.Finally,according to the data segmentation point and the dynamic range segmentation point,the bi-histogram equalization is performed,and the final enhancement result is obtained.Experimental results show that compared with other algorithms,our method has wide application range,strong detail retention ability and natural effect.(2)A mixed model based global histogram equalization algorithm is proposed.Firstly,according to the proposed mixed model,the optimal truncation threshold is found.Then the histogram is truncated according to the optimal truncation threshold,and the truncated part is evenly distributed to the effective gray level of the histogram.After that,the postprocessing of the histogram is carried out.Finally,traditional histogram equalization is performed to obtain the final enhancement result.The experimental results show that our method has appropriate contrast and a widely application compared with other algorithms.And it can be used for real-time video enhancement.(3)Finally,a universal video real-time enhancement system based on video acquisition card is developed on the Visual Studio platform.The system receives HDMI video signal through the video acquisition card,and then displays the enhanced image on the system software interface after real-time processing.The test results show that the system realizes the real-time enhancement processing of 60 frames 1080*1920 video in many scenes,and the enhancement effect is natural.Since the parameters are self-adaptive,there is no need for manual operation and it has practical value.In conclusion,aiming at the shortcomings of the existing improved histogram equalization algorithms,this paper proposes two improved histogram equalization algorithms based on the model one after another,and implements one of them in engineering.