Cotton Color Distributions in the CIE L*a*b* System
Cotton Color Distributions in the CIE L*a*b* System
复制标题
CIE L*a*b* 系统中的棉花颜色分布
DOI：
--
复制
发表时间：
2001
期刊：
影响因子：
--
通讯作者：
M. Watson
中科院分区：
文献类型：
--
作者：
Bugao Xu;Y. Huang;M. Watson
关于CIE L*a*B* 中的棉花颜色分布的信息对于开发用于颜色分级的新色度技术是至关重要的，因为CIE L*a*B* 是比棉花工业中当前使用的Rda B系统更标准化的系统。本研究的目的是研究棉花颜色在CIE L*a*B* 中的分布，并探索单个颜色等级（类）的概率密度函数模型。根据美国农业部高地棉物理标准的试验数据和1999年的棉花样品，棉花颜色在CIE L*a*B* 空间中是自然可分的，颜色等级的概率分布呈正态分布，相邻两个等级的概率分布有一定程度的重叠。这项工作演示了如何构建一个多元正态分布的颜色等级的基础上，在该等级的样本的统计数据，以及如何确定两个相邻等级之间的硬决策边界，最大限度地减少错误分类。颜色分布模型有助于建立更有效的颜色分级方法。
Information about cotton color distributions in the CIE L*a*b* is critical to developing new chromatic technology for color grading, since the CIE L*a*b* is a more standardized system than the Rdab system currently used in the cotton industry. The objectives of this study are to examine the distributions of cotton colors in the CIE L*a*b* and to explore models of the probability density functions of individual color grades (classes). Based on the experimental data of the USDA physical standards for Upland cotton and samples from the 1999 crop, cotton colors are naturally separable in the CIE L*a*b* space, the probability distributions of color grades are normal, and the distributions of two adjacent grades overlap to some extent. This work demonstrates how to construct a multivariate normal distribution for a color grade based on the statistical data of samples in that grade, and how to determine a hard decision boundary between two adjacent grades that minimizes misclassifications. The color distribution models are helpful for establishing more effective classification methods for color grading systems.