Research And Algorithm Implementation Of HDR Display

HDR(High Dynamic Range)images have high dynamic range and high contrast.The current general video source of the dynamic range is not wide enough,and the traditional display can only show the contrast dynamic range of 1000:1.The human eye can feel the brightness range of about 10-3?106nit,can feel the instantaneous contrast range of up to 10000:1.HDR technology can achieve 0.01nit?1000nit dynamic display range,so that the image darker,bright brighter,the details stronger.HDR display key technologies include digital dynamic range enhancement technology,local dimming technology and peak drive technology.Because most of the images on the market are SDR(Standard Dynamic Range),this paper designs an algorithm based on histogram equalization to improve the contrast of SDR images,so that SDR images can get performance close to HDR.1.This paper studies a number of techniques to improve the contrast,which focuses on the histogram equalization technology.The reason that histogram equalization can improve contrast but easy to over-increase is analyzed.In order to ensure picture quality,histogram equalization techniques need to be limited.2.In order to improve the contrast of digital images,a histogram equalization algorithm using two modulation factors is proposed for digital dynamic enhancement.Histogram equalization can improve the image contrast,the global histogram equalization structure is simple,but easy to make the image enhancement and loss of detail,although the local histogram can highlight the details but need to calculate the mapping function for each point.For this article,an Algorithm for Combining Global and Local Histogram Equalization is proposed.The algorithm uses only a mapping function that is equalized by the traditional global histogram,and then divides the image into multiple partitions and defines the local limiting parameters and the point decision factors,which reflect the local information and the local flatness respectively.The two local modulation factors modulate the mapping of each point.The algorithm uses bilinear interpolation to smooth the partition local limit parameters.In order to evaluate the effect of the algorithm,this paper uses the absolute mean brightness error(AMBE)and discrete entropy(DE).The evaluation result is that the algorithm has better ability to maintain average brightness and can improve the discrete entropy of the data,which can improve the contrast and avoid image degradation.3.In order to realize the algorithm in the real-time system,the algorithm has been designed and transplanted to the FPGA(Field programmable gate array).The results show that compared with the global histogram equalization,the enhancement effect is better,the adaptability is stronger,compared with the local histogram equalization,the use of resources less.By the algorithm design and FPGA implement,in this paper,the low dynamic range image is successfully converted to a high dynamic range image in real time,so that the general video source can also have the performance effect close to the HDR source.