基于近红外光谱和K线图的烟丝总糖实时监测方法  

K-line tracking method of total sugar changes in cut tobacco

作　　者：杜国荣 马莉 马雁军 刘德水 史素娟 矫海楠 黄越 DU Guorong;MA Li;MA Yanjun;LIU Deshui;SHI Sujuan;JIAO Hainan;HUANG Yue(College of Food Science and Nutritional Engineering,China Agricultural University,Beijing100083,China;Technology Center,Shanghai Tobacco Group Beijing Cigarette Factory Co.,Ltd.,Beijing 101121,China;Beijing Life Science Academy(BLSA),Beijing 102211,China)

机构地区：[1]中国农业大学食品科学与营养工程学院,北京100083 [2]上海烟草集团有限公司技术中心北京工作站,北京101121 [3]北京生命科技研究院有限公司,北京102211

出　　处：《中国烟草学报》2023年第5期20-26,共7页Acta Tabacaria Sinica

基　　金：北京市自然科学基金项目(No.8222070);兴化健康食品产业研究基金(No.201905);上海烟草集团有限责任公司科技项目(No.K2015-2-017p);上海烟草集团北京卷烟厂有限公司科技项目(No.TP2019-C1)。

摘　　要：【背景和目的】针对传统生产过程监控中对特定化学成分的变化趋势往往无法实时获取和判断的问题,提出一种基于金融领域中K线图计算和趋势判断的质量监控方法。【方法】通过颜色趋势和长短均线策略实现对批次内、批次间结果的比对分析并预测未来批次走势,并与常用的误差棒图和箱线图进行对比分析。【结果】牌号A和B的烟丝总糖趋势判断准确率最高分别为75.00%和84.62%,采用K线图能判断批次前期、后期的物料化学成分变化。【结论】该方法具有较好的直观性和时效性,对同类产品的其他生产指标监控具有良好的拓展应用价值。The change of total sugar content in cut tobacco in the cigarettes production is one of the main indicators affecting the stability of the product.Aiming at the problem that the change trend of specific chemical components cannot be obtained and judged in real time in traditional production process monitoring,this study proposed a quality monitoring method based on K-line chart calculation and trend judgment from the financial field.First,based on the data of cut tobacco obtained by online near-infrared spectroscopy,the total sugar content of 525 batches from two brands was analyzed before and after the short-term moving average or the long-term moving average.Comparative analysis of intra-batch and inter-batch results and prediction of future batch trends were successfully realized.Results showed that the highest accuracy rates of judging the trend of total sugar in cut tobacco of brands A and B were 75.00%and 84.62%,respectively.Subsequently,the commonly used error bar chart and boxplot were compared with the K-line analysis and the results revealed that the K-line combined with the long and short moving averages can accurately predict the trend of future batches better.This approach has good intuition and timeliness,and has a good expansion and application value for the monitoring of other indicators of similar products.

关 键 词：烟丝 K线图 过程监控 总糖含量 近红外光谱 

分 类 号：TS47[农业科学—烟草工业] O657.33[理学—分析化学]