文章编号：1004－5422(2022)01－0052－06

DOI:10.3969/j.issn.1004-5422.2022.01.009

基于混合颜色模型的茶叶多级分类技术

周　敬,张建伟,张光龙,周　强

（成都大学 电子信息与电气工程学院，四川 成都 610106）

摘　要：针对目前茶叶色泽质量判断与分级容易受人为主观因素及心理因素影响的现状，提出用计算机视觉技术图像中的RGB和 HSI混合颜色模型，并以绿色分量、色度和饱和度作为颜色特征向量，通过贝叶斯决策去除劣质茶叶，再通过K均质聚类分为第1等级、第2等级.以某干茶为实验对象，针对茶叶的色泽特征，以绿色分量判别色泽优劣等级，再以色度与饱和度相结合进行分析，构建特征向量函数并进行聚类二次分级.实验结果表明，所提出的方法对该类茶叶色泽质量分类识别的准确率达92.5%.

关键词：计算机视觉；混合颜色模型；分类；K均值

中图分类号：TP391.41;TS272.7   文献标志码：A

Tea Classification Technology Based on Mixed Color Space

ZHOU Jing,ZHANG Jianwei,ZHANG Guanglong,ZHOU Qiang

（School of Electronic Information and Electrical Engineering，Chengdu University,Chengdu 610106,China）

Abstract：At present,most of the evaluation methods of tea color are easily influenced by psychological and subjective factors.Therefore,this paper,based on RGB and HSI of computer vision,firstly uses mixed color space to identify the grades of tea by analyzing the color of tea with green component,chromaticity and saturation as color feature vectors.Secondly,bad tea is eliminated by Bayes Decision,which is classified into excellent and general categories by K means clustering.Taking the dry Queshe tea as an example,according to the color characteristics of tea,the good and bad grades are judged by green component,and then analyzed by combining chromaticity with saturation,and the feature vector function is constructed and the clustering algorithm is used.The experinent results show that the accuracy of this method is as high as 92.5%.

Key words：computer vision;mixed color space;classification;K means