摘要
彩色显示器是传达信息的主要设备之一,但相同的RGB颜色值在不同的显示设备上视觉效果往往有差异,导致颜色传递不准确,因此有必要建立一个色度特征化方法,使RGB颜色转变到设备无关的CIELAB颜色空间,以提高颜色表现的准确性。为解决该难题,借助多项式描述RGB与CIELAB之间的非线性关系;建立了显示器设备的RGB到CIELAB的建模和精度测试色靶;建模过程中,引入混沌序列,对粒子群优化方法进行改善,基于混沌粒子群优化方法以均方误差为适应度函数估计出各个多项式的系数,从而完成建模过程。将所提方法与最小二乘和基于MCMC的贝叶斯估计方法进行了比较,精度基本接近,但耗时更少。结果表明:该模型针对测试色靶的平均色差较小,得到了较好的结果,是一种有效的显示设备色度特征化方法。
Color display is one of the main equipment of conveying information. However,color visual effect often has differences in different devices with the same RGB value. Thus a color characteristic method is needed to transform the device-dependent RGB color space to the device-independent CIELAB color space. To solve this problem,the polynomial regression is used to describe the nonlinear relationship between RGB and CIELAB.Then,the modeling color targets and performance testing color targets are set up. During the modeling process,chaotic sequence is introduced to improve the particle swarm optimization method. Mean square error( MSE) is taken as the fitness function,and the polynomial coefficients are estimated based on the chaotic particle swarm optimization method. The proposed method is compared with the least squares method and Markov-Chain-MonteCarlo( MCMC) based Bayesian estimation method. The results show that the accuracies of these three methods are nearly the same and the chaotic particle swarm optimization( CPSO) method is less time-consuming. The experimental results also show that the average color difference of the established model is little,and the proposed method is an effective method for display device colorimetric characterization.
出处
《机械科学与技术》
CSCD
北大核心
2015年第1期124-130,共7页
Mechanical Science and Technology for Aerospace Engineering
基金
国家自然科学基金项目(61471295
61172123)
陕西省科学技术研究发展计划项目(2013K07-18)
陕西省教育厅科学研究计划项目(14JK1524)
陕西省自然科学基础研究计划项目(2014JM2-6111)资助
关键词
色度特征化
颜色空间转换
混沌粒子群优化
色差
多项式回归
chaos theory
chaotic particle swarm optimization
color difference
color space transformation
colorimetric characterization
convergence of numerical methods
display devices
efficiency
estimation
experiments
global optimization
least squares approximations
Markov processes
mean square error
Monte Carlo methods
particle swarm optimization(PSO)
polynomials
regression analysis
