【作者】 王义峰；

【导师】 曾平；

【作者基本信息】 西安电子科技大学 ， 计算机应用技术， 2008， 博士

【摘要】 随着计算机科学及色彩输入/输出技术的发展,彩色图像作为信息载体,在印刷、影像、广告、影视、电子商务、数字娱乐等诸多领域得到了越来越广泛的应用,人们对色彩再现质量的要求也越来越高。然而受成像机理、色彩空间、元器件性能、耗材特性、机械加工精度等因素影响,输入/输出设备的彩色特性差异十分显著,这使得彩色图像跨介质、跨平台一致再现极为困难。为实现“所见即所得”(WYSIWYG)的目标,必须对颜色进行测量、控制和管理,于是形成了多种色彩管理体系结构,其中最著名的当属国际色彩联盟(ICC)提出的色彩管理规范。ICC色彩管理引入设备无关的色彩空间(通常为CIELab或CIEXYZ)统一描述设备彩色特性,通过色域匹配及设备输入/输出关系非线性校正,建立设备相关色彩空间(RGB、CMYK等)与设备无关色彩空间之间的映射关系,实现彩色信息从输入设备色彩空间到设备无关色彩空间,再到输出设备色彩空间的准确转换。针对现有色彩管理体系不具备自适应渲染功能、无法获得最佳图像再现的不足,本文重点研究色彩管理体系结构的改进及设备色彩特性化、设备色域描述、自适应色域匹配等关键技术,并根据理论研究成果,设计了一套色彩管理应用支撑软件。从最初的封闭式色彩管理体系到ICC开放式体系、PostScript体系、sRGB体系、基于色外观的体系以及光谱再现体系,色彩管理在发展中绽放出多种形式。由设备彩色特性描述文件(Profile)和色彩管理组件(CMM)组成的ICC色彩管理已成为彩色工业的通用标准。理论分析和实际应用中发现,目前基于ICC规范的色彩管理系统不能针对图像进行自适应渲染,因而无法获得最佳的色彩再现。针对该问题,提出一种基于ICC规范扩展的色彩管理体系结构,其核心是开发具有自适应色域匹配功能的CMM,针对图像特征进行最佳渲然。虽然与重Profile、轻CMM的传统ICC色彩管理体系侧重点相反,两系统外特性上却完全相同。扫描输入、显示和打印输出设备中,以打印机彩色非线性最复杂。因此本文以CMYK喷墨打印机为主要研究对象,首先进行设备彩色特性提取,重点是打印机的色域描述和色彩校正。实验发现打印机墨水呈色存在显著的非线性特征,于是引入了B样条描述非线性,在此基础上提出一种基于样条张量积的打印色彩校正算法,获得相当校正精度的条件下,所需打印测量的样本数较常规插值算法有大幅下降。提出一种基于设备色域表面采样和样条插值的色域描述算法,计算所得色域体积、形状与真实设备色域极为接近。两种色彩特性提取算法普适性好,可推广到扫描、显示及超四色打印设备。针对超四色喷墨打印机,提出一种基于色域划分的色彩校正算法,降低了颜料空间的维数,消除了颜色冗余。设备间色域形状、大小及位置差异是彩色跨介质一致再现的难题。因此色域匹配成为色彩管理系统中的一项核心技术,既要调整彩图颜色使之落入再现设备色域内,又要尽量保持原始图像的视觉效果。针对当前ICC体系中色域匹配算法与图像无关,难以充分发挥设备再现能力的不足,提出两种图像特征相关的自适应色域匹配算法。基于图像类别的匹配首先提取图像特征并分类,根据图像类别确定合适的匹配策略。基于视觉评价模型的算法考虑了人眼对再现图像的主观评价,同时保证图像中整体色彩和边缘细节的损失达到最小。针对国内目前缺乏具有自主知识产权色彩管理应用支撑软件的现状,设计并实现了基于ICC扩展的色彩管理系统。该系统由彩色设备Profile生成器和自适应色彩管理组件CMM构成。Profile生成器应用了设备彩色特性提取技术的研究成果,CMM则集成了自适应色域匹配功能。另外CMM设计完全遵循Windows图像色彩管理规范,最终以Windows第三方CMM的形式在操作系统级别进行色彩管理。实际应用表明,自适应色彩渲染要优于Windows自带色彩管理功能。更多还原

【Abstract】 With the development of computer science and color input-output technology,color image as a media of information is applied more and more widely in such areas as press, imaging, advertisement, movie, e-business and digital entertainment et al and people raise a higher request for the quality of color reproduction. While factors like imaging mechanism, color space, apparatus capability, material peculiarity and machining precision make the IO devices’color characteristics quite different. So it is hard to achieve a platform or media-cross conformable color reproduction. To accomplish WYSIWYG (what you see is what you get) people have to measure, control and manage color. As a result, many color management systems(CMS) are developed which is far from each other. Among those, ICC color management system is the most famous one. ICC color management system introduces device independent color space (usu. CIELab or CIEXYZ) to describe the color characteristic of all the devices in a uniform space. Through gamut mapping and nonlinear calibration on the input/output characteristic of the device, this system creates a mapping between device dependent color space (RGB or CMYK) and device independent color space, which helps to transform the color information from the input device color space to the device independent color space, and finally, to the output device color space. Since color management system frameworks available can not render images adaptively to get a best reproduction, this paper focus on color management framework improvement and such core technologies as color characterization, color gamut description and gamut mapping is discussed. Then on the basis of the theoretic research, an application software of color management based on extended ICC specification is developed.Many shapes of CMS have come forth: close-looped, ICC-based, PostScript, sRGB-based, color appearance involved as well as spectrum based reproduction system. So far, ICC which is made of profile and color management module is in fact the standard of color industry. While theoretical analysis and applications show that ICC can not render images adaptively to get a best reproduction. So an extended framework whose kernel part is a CMM with adaptive color gamut mapping ability is proposed, the CMM will render color corresponding to image characteristics. Though on the contrary that ICC lays particular stress on profile than CMM, the proposed CMS has the same interface of ICC color management system. Among the computer peripherals, printer is the device that has most serious color nonlinearity. Focus on CMYK ink jet printers, this paper researches on color characteristic calculation with emphasis on gamut description and color calibration. Experiments show that ink colorants are quite non-linear to produce colors, so B-spline is introduced to account for the nonlinearity. Then a spine tensor-product based color calibration algorithm and a surface data sampling with spline interpolation gamut calculation method are given. Experiments show that the proposed calibration method can get an equivalent precision with much less measure samples, and the gamut description gained by the proposed algorithm is quite close to the real one. At the last of this chapter, we propose a multi-channel ink jet printer calibration algorithm by printer color gamut partition and subspace characterization. Partition reduces the data dimension and eliminates color redundancy.Different devices have different color gamuts, and gamut mapping which can map all colors to the destination gamut must be done in advance when color information is to be reproduced. In view of the limitation of the present ICC color management system, a further study is made on adaptive color gamut mapping .Two image content relevant gamut mapping algorithms are proposed. The first one classify images into different groups and each group assign a proper matching method; the second algorithm is based on human vision system, which can preserve image colors as well as image gradients as much as possible.In view of the situation that so far there is few native color management application software, we design and accomplish the ICC-extended CMS. It is composed of a profile creator and a CMM with adaptive color management. The profile creator adopts all research results of the device characterization and the CMM aggregate the adaptive gamut mapping technique. Besides, the CMM conforms to windows image color management specification and it is a 3rd-part CMM working on the operation system level. Experiments show that the adaptive color rendering is superior to the windows default color management.更多还原