基于图像分析的液晶显示动态调光问题研究

发布时间：2018-02-11 11:00

本文关键词： 全局调光主灰度图像分类分区调光多重分区背光分配镜头特征　出处：《合肥工业大学》2016年博士论文　论文类型：学位论文

【摘要】：液晶显示器经过多年来的发展已经成为电子通信、家用电器、航空航天、工业控制和汽车电子等领域的主流显示设备。液晶显示器作为一种非自主发光的显示器件需要使用背光模组作为它的光源。背光模组主要由LED及光学膜系构成,依据LED放置位置不同分为直下式和侧出式背光模组。一般液晶显示器的背光模组输出亮度恒定,这会导致显示器功耗过高且存在漏光问题。所以动态调光技术被提出,该技术是根据图像内容动态控制背光亮度并对液晶像素进行补偿,能够有效降低能耗并提高对比度。动态调光技术依据对背光模组的控制方式不同可以分为全局调光技术和分区调光技术。本文旨在深入分析现有的动态调光技术,研究动态调光的核心问题。依据数字视频处理的相关原理,探索视频分割方法与动态调光技术的结合运用；使用图像特征提取和图像增强等技术,改进动态调光算法涉及的背光提取和图像补偿方法；完成全局动态调光显示系统和分区动态调光显示系统的开发工作。主要的研究工作包括：1)综合分析了现有背光提取方法的优缺点,建立了一套用于动态调光算法评价的指标体系。使用统计分析方法建立图像特征与合理背光值之间的关系,从而提出了一种基于直方图和主灰度背光提取算法。2)用于动态调光的S型曲线调节参数对于不同类型的图像适应性不够好,所以使用图像分类和人脸图像判别等算法分析图像类型,依据图像类型调节S型曲线参数,以便达到更好的显示效果并降低能耗。同时针对S型曲线参数快速变化导致图像闪烁的问题进行修正,引入a值平滑机制来有效地减少图像闪烁。3)为了适应不同背光结构和分区方式提出了一种多重分区动态调光算法。通过采用多重分区划分图像确定每个分区的背光值。针对侧出式分区背光算法的背光分配问题,使用模拟退火算法基于能量消耗最低原则求取近似最优的背光值分配方案。4)提出一种基于视频镜头特征的动态调光方法。该算法使用亮度、主颜色直方图和颜色方差参数来分割视频,确定每个镜头的视频类型,依据视频类型选择合适的调光参数。5)依据本论文的算法研究成果设计了三款动态调光显示器样机并进行了显示效果测试和节能率测试。
[Abstract]:After years of development, LCD has become electronic communication, home appliances, aerospace, Liquid crystal display (LCD), as an involuntary luminous display device, needs to use backlight module as its light source. The backlight module is mainly composed of LED and optical film system. According to the position of LED, it can be divided into straight down and side out backlight modules. The output brightness of LCD backlight module is constant, which will lead to high power consumption and light leakage. So dynamic dimming technology has been proposed. The technology is to dynamically control the backlight brightness according to the image content and compensate the LCD pixels. Dynamic dimming technology can be divided into global dimming technology and zonal dimming technology according to the control mode of backlight module. The purpose of this paper is to deeply analyze the existing dynamic dimming technology. This paper studies the core problems of dynamic dimming. According to the related principles of digital video processing, it explores the combination of video segmentation method and dynamic dimming technology, uses image feature extraction and image enhancement techniques, etc. Improving the backlight extraction and image compensation method of dynamic dimming algorithm; The development of global dynamic dimming display system and zonal dynamic dimming display system is completed. The main research work includes: 1) the advantages and disadvantages of existing backlight extraction methods are comprehensively analyzed. A set of index system for dynamic dimming algorithm evaluation is established, and the relationship between image features and reasonable backlight value is established by using statistical analysis method. Therefore, an algorithm based on histogram and main gray level backlight extraction algorithm. 2) for dynamic dimming, S-shape curve adjustment parameters are not good enough for different types of images. Therefore, image classification and face image discrimination algorithms are used to analyze image types, and S-type curve parameters are adjusted according to image types. In order to achieve better display effect and reduce energy consumption. At the same time, the problem of image flicker caused by the rapid change of S-shape curve parameters is corrected. In order to adapt to different backlight structure and partition mode, a dynamic dimming algorithm for multiple partitions is proposed. The back of each partition is determined by using multi-multiple partition images. For the backlight assignment problem of the side-out partition backlight algorithm, A dynamic dimming method based on video shot features is proposed by using simulated annealing algorithm to obtain approximately optimal backlight value allocation scheme based on the principle of minimum energy consumption. The main color histogram and color variance parameters are used to segment the video and determine the video type of each shot. According to the algorithm of this paper, three dynamic dimming display prototypes are designed, and the display effect and energy saving rate are tested.
【学位授予单位】：合肥工业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP391.41;TN873.93

【相似文献】