面向节能的图像颜色优化研究硕士研究生学位论文bewt1g50.docVIP

硕士学位论文 面向节能的图像颜色优化研究 摘要 在目前的桌面电脑和移动设备当中，显示器所占用的电能功率消耗比例分别达38%和50%，因此节能显示技术的研究，特别是在延长移动设备电池持续供电时间等方面有着十分迫切的需求。从技术上讲，显示内容的节能显示主要可以从硬件和软件方面进行。 在对显示器节能和图像处理相关方法进行了综述的基础上，本文从软件处理角度出发，围绕图像的节能显示优化设计的问题展开了研究，并提出了一种基于节能的图像自适应亮度优化方法。该方法首先用双边滤波对图像进行去噪，然后通过Sobel边界检测算子从图像中获取显著边界作为图像主要特征，随后构造特征重要性场，用于指导图像的特征增强，并实现自适应亮度优化。该方法在应用于图像亮度降低时，可以在保持特征分辨能力的同时获得节能显示的效果。继而，本文描述了一个实现上述图像处理方法的软件原型，并提供基本的图像处理操作，方便用户对方法的参数调节进行预览，并对处理结果进行分析比较。 最后，本文将上述方法应用于一般图像进行了实例测试，并做了简单的用户研究，结果表明该方法能够在相同节能效率下更好地保持图像中的感知特征。 关键词：节能，颜色优化，视觉感知，边界检测 Abstract In the commodity desktop computers and mobile devices, the monitor consumes relatively 38%~50% electrical energy of the total consumption. Thus, designing an energy saving scheme is very essential, especially in extending power supply of the mobile device battery. Technically, energy saving display can be achieved by both hardware and software solutions. In this thesis, we conduct a study on image brightness optimization problems for energy saving display based on the survey on energy-saving schemes and image edge detection. In particular, we propose an adaptive dimming approach for display energy saving. In this approach, bilateral filtering is performed upon the original image firstly to smooth the noise while preserving evident features. Then we construct a saliency map by extracting edge features from the image using Sobel operator. The saliency map describes the region of interest which plays an important role in users’ understanding of an image. In the optimizing process, we perform a saliency guided feature enhancement by using the saliency map. With the proposed approach, we can generate a dimmed image which can be displayed with less energy consumption. Additionally, we describe a software prototype which implements our method and some elemental image operations. Users can preview the intermediate results while they are tuning the relevant parameters, as well as analysis and comparing of the results. We apply our approach on several applications, in