Discussion on the development of domestic (CN) Colorschemer studio

本文叙述了研制适合我国国情的电子测配色软件过程中解决的技术问题。软件系统总体设计，配色运算原理，数值运算处理，引入染料特性曲线，搡作倌息中文化、数据结构的开放性，数据显示宜观化，影响配色精度因索的消除及染料色泽波动的适应性。同时，将我们的软件与进口同类产品进行了比较，提出了使软件适合我国国情的一种设计思路。

—、概述

随着我国改革开放、建设社会主。义市场经济的步伐不断加快，用髙科技改造和装备纺织这一传统工业已势在必行.一个面对市场竞争激烈、流行色周期日趋缩短及服装色泽多变的印染厂，要使自己立于不败之地，就需要以客户的需求为主导，以小批量、多品种的方式组织生产.显然，传统的手工方式已越来越难以相适应.作为现代高新技术代表的电子计算机技术具有大规模的存储手段，髙速运算能力和髙度的智能，并且借助于色度学的理论，能够在极短的时间内，对许多数据进行大量的处理和运算，提供多个合适的染色处方预告.电子计算机配色主要有三种方式，色号归档检索，反射光谱匹配和三刺激值匹配.其中三刺激值匹配法应用最为广泛.我们研制的测色配色系统提供了按色号归档检索和三刺激值匹配分别进行配色的功能，可供使用者选择.应,用实践证明了软件计算精度高，运算速度较快，使用方便，尤其是在显示信息中文化、操作直观矣掌握上.更受到应用厂家的欢迎.

二.普系统设计的基本问題

1。电子计算机软件总体设计由于无法借鉴进口同类产品，我们完全依靠自己的力量来开发软件.电脑配色有两个显著的特点，其一是数据的庞大性和复杂性，在使用时要求能够迅速调出有关的信息.我们定反射率数据位数为六位，反射率曲线取十六个不同波长点的值.因此，存储98个坯布数据需要98X16X8个字节.如果每个坯布对应30个染料，每种染料可存储10档不同深浅色样.則存储染料数据需要2940x10xl6x6个字节.其二是配色运算处理速度的实时性，要求能在很短时间内在全部可能的染料组合中求出所有合理的处方/进口软件的速度一般為:在IBMPC/XT级微机上算一个组‘合约需几十秒时间.

由于每一种计算机语言都有k待定应用方向，大多在某一方面功能较为突出，因此选用单一语言很难兼顾两方面要求.我们采用了将dBASEm编译和另一种编译语言混合编程的方法.前者具有很强的数据分类存储和快速査找能力，而后者具有运算速度快，作图功能强的特点。

三刺激值配色系统的主要数据集有兰个，—是染料数据：二是坯布数据I三是标样即客户来样数据.各程序之间仅逋过这些数据发生联系，这样构造的系统具有极强的开放性。便于不断增加的应用功能的程序的兼容性.

2.配色运算设计的色度学丨衣据电脑配色所用的计算公式也是需要靠我们自己设计.

整个测配色系统中，配色软件是中心部分，占据着重要的位置.我们所研制的紀色软件，采用的是Allen三刺激值匹配矩阵计算法，它的数学原理为一组联立的全微分方程：

3.数值运算处理过程中的问题

色度学从理论上为配色运算的实现提供了According to，但在具体的运算处理中，还存在一些实际问题需要加以解决.例如：迭代运算退出的控制，求解快速逼近.在设计配色软件时，我们设置了多个迭代运算退出的条件.其中之一是According to自动控制的稳定性理论，自动判别运算的发散收敛情况.如同一个控制系统的状态在接近给定要求时会产生先逼近后发散的情况，配色迭代也会发生这种情况.我们的软件自&够在很好的接近点停止运算并保留结果.另一方面，迭代运算本身的质量在很大程度上取决于染料浓度每次迭代时增量(或称步长〉Ac,的设定.取得太小，则迭代次数增加，逼近很好的值的过程太长：反之，则有可能因求解祖糙而丢失准确的结果.经过反复试验和比较，我们总结出一个能够兼顾的值.以上几点措施经实践..证明是行之有效的。我们做过与进口ACS配色系统软件的对比试验.在完全采用进口仪器测得的基础数据为素材的条件下，分别对五只来样进行配色，结果证明无论是在速度还是在精度方面，国产的软件均达到与进口软件相同的水平（一次成功率达到70%).

4.染料特性曲线的获得

配色运算是以染枓、坯布基础数据为基本According to进行的，因此，基础数据置备工作是保证整个系统正确运行的一个关口。由于存在着打样时实验条件方面的误差及测色仪的测量误差，因此，尽管从理论上说，染色布样在各个频率点的吸收因子值与所用染料浓度成正比，并且当浓度大到一定值时，值趋于饱和.实际做出的各浓度值与K/S值的相关分布总有些波动.因此，我们设计了基础染料数据的校核功能.每当送入一个染料后，要先做校梭.由计算机采用分段拟合方式对浓度t/T/S值相关分布进行拟合.According to拟合离M最小的原则，拟合自动取一次或二次曲线.拟合的结果再直接显示于屏幕上，由操作人员对个别跳跃点进行修改，然后再由计算机拟合.直至使用者满意为止。在配色时，是直接用拟合曲线来求某一浓度所对应的K/S值.实现了比较专业的运算方法.

In many imported software, we judge according to its operation mode and sequence, and use the two-point interpolation method to find out the redundant /S when color matching. value, so it is impossible to overcome the error caused by the jump of the basic number of dyes. Obviously, the performance of our software in this area is much better than that of imported products.

3. Enhancement of software performance to suit national conditions

1. Operation information Chinese culture

Imported software is explained in foreign languages, and there are not many prompts, which brings great inconvenience to domestic use and increases the difficulty of popularization and application. We believe that only color matching and proofing personnel directly participate in the selection and application of computers. It has vitality. Therefore, when designing domestic software, it is necessary to remind and explain all Chinese culture from beginning to end. In terms of complete information prompts, try to include the original data, current processing status, functions, result data, and the next step of the operator. What kind of information is all displayed on the screen, and its completeness has exceeded the import system of various brands.

2. openness of data structure

Due to commercial considerations, foreign and some domestic software are generally completely sealed, and users have no way of knowing. The main data of the software developed by us is stored in the database file of dBASEIII. Since dBASEIII is a very popular database system, these data are easy to obtain This provides a feasible means for users or non-original development units to expand and adjust the software system, which is undoubtedly easier to be accepted by application manufacturers, so that they can not only use it passively, but can adapt it according to changes and their own needs. The situation participates in the operation process of the system itself.

3. Intuitive display of processing results

There are a series of steps and a certain complexity in operating the computer color measurement and matching system. If the operator can understand the work at a glance, it will help reduce the psychological pressure caused by the lack of understanding of the internal work of the computer.. We are designing When using computer software, try to use the intelligence of the computer to present the complex work it does to the operator in an as simple and intuitive way as possible. For example, the color matching function of the dye curve diagram: after the gray cloth used, the dye and its negative concentration are given Then test the incoming dye samples, and you can get the reflectance curve of each dye and gray fabric on the screen, the incoming dye samples and the color matching prediction curve. Therefore, the dyeing workers can not only make some tentative color matching, but also calibrate For example, in the color difference measurement function, we not only display the color difference value obtained by various calculation formulas, but also display the comparison curve and the distribution of the measurement points of the two color samples in the AB bridge and ^ coordinate system. These aspects are the salient features of domestic software.

4. Elimination of factors affecting color matching accuracy and adaptability to fluctuations in dyeing color

The AT/S value of the color matching result is a linear superposition of the K/S value of each dye, which is a satisfactory state. In practical applications, there are many factors that will affect its correctness. For example: when color matching The color saturation of the dye is obviously much higher than the value of the single dye basic data, the difference in the coloring speed of the dye when the color is mixed, and the quality and stability of the dye are not high, etc., which will affect the accuracy of the result.

There are many different methods for eliminating the factors that affect the accuracy of color matching. A traditional method is to conduct many experiments and summarize a set of correction coefficients for each factor. During the color matching process, these factors are one by one. This method belongs to feed-forward control, which is feasible in theory, but there are many problems in practice. First, when calculating the size of a single factor, it cannot completely make other factors ineffective. At the same time, by sub-real Eyelids and Yundi bring new errors.