中药饮片性状质量智能检测关键技术研究  

Research on Intelligent Detection Technology of Appearance Quality of Traditional Chinese Medicine

作　　者：周明[1] 周金海[2] 张燕群 段念林 Zhou Ming;Zhou Jinhai;Zhang Yanqun;Duan Nianlin(School of Traditional Chinese Medicine and School of Integrated Chinese and Western Medicine,Nanjing University of Chinese Medicine,Nanjing 210023,China;School of Artificial Intelligence and Information Technology,Nanjing University of Chinese Medicine,Nanjing 210023,China;Hangzhou Haishan Pharmaceutical Equipment Co.,Ltd,Hangzhou 311401,China)

机构地区：[1]南京中医药大学中医学院·中西医结合学院,南京210023 [2]南京中医药大学人工智能与信息技术学院,南京210023 [3]杭州海善制药设备股份有限公司,杭州311401

出　　处：《世界科学技术-中医药现代化》2023年第5期1580-1589,共10页Modernization of Traditional Chinese Medicine and Materia Medica-World Science and Technology

基　　金：国家科学技术部国家重点研发计划中医药现代化研究重点专项(2019YFC1711500):中药材净切制关键技术与智能设备研究及应用,负责人:曹晖

摘　　要：目的基于人工智能之计算机视觉技术,构建客观化的中药饮片性状质量智能检测模式,以高科技应用促使中药饮片企业提质增效。方法以大黄饮片为例,利用开源的OpenCV视觉库,设计智能的中药饮片性状质量检测方法。步骤:运用斑点检测法检测中药饮片表面斑痕,对饮片进行初步筛选;轮廓提取结合凸包检测,从片形大小角度对饮片进行等级划分;获取最小外接矩形和凸缺陷检测,对异形饮片进行排除;提取饮片颜色特征数据,经机器学习方法建模,从色泽角度对中药饮片进行优劣判别。结果斑点、大小、异形检测均可有效排除不合格饮片,色泽检测模型对大黄饮片质量判别的准确率达84.00%。结论中药饮片性状质量智能检测关键技术的实现,可用于弥补传统人工分拣质量与效率的不足,为中药饮片行业全线产品在线质量智能检测系统研究奠定了基础,或将有力推进中药饮片生产现代化进程。Objective Based on the computer vision technology of artificial intelligence,an objective intelligent detection mode for the appearance quality of Chinese herbal pieces was constructed,so as to promote the traditional Chinese medicine enterprises to improve the product quality and increase the benefits with the application of high technology.Methods Taking Rheum Palmatum as an example,we used the OpenCV visual library to build an intelligent inspection method for the appearance quality of decoction pieces.The specific steps were as follows:Firstly,the spot detection method was used to detect the surface markings of the pieces,and it was the preliminary screening of the quality of the pieces.Secondly,through contour extraction and convex hull detection,the decoction pieces were classified according to the areas.Then,the minimum enclosing rectangle and convex defect detection methods were used to discriminate and eliminate irregular pieces.Finally,machine learning modeling was used to identify decoction pieces from the perspective of color.Results The spot detection,size detection and shape detection could effectively eliminate unqualified pieces.The accuracy of color detection in judging the quality of Rhubarb slices was 84.00%.Conclusion The realization of the key technology of intelligent detection of the appearance quality of Chinese herbal pieces could be used to make up for the deficiency of the quality and efficiency of traditional manual sorting,and laid a foundation for the research of the online quality intelligent detection system of the whole line products in the Chinese herbal medicine industry,or would effectively promote the modernization of the production of Chinese herbal pieces.

关 键 词：中药饮片 性状质量检测 计算机视觉 OpenCV视觉库 

分 类 号：R282.5[医药卫生—中药学]