摘要
面团的发酵过程受到面粉种类、温度、加水量、酵母量等诸多因素的干扰,通常须人工经验掌控,耗时耗力,难以掌控。采用机器视觉技术,搭建了图像采集试验台架,抓取了面团发酵过程的图像,提取了面团图像的5个特征,确定了面团发酵的时域过程:充满区-上涨区-停滞区,使用决策树算法对数据集进行了筛选和训练,获得了用户友好、逻辑清晰的面团发酵过程可视化树。十折交叉验证结果表明,基于决策树算法的面团发酵过程分类树的精度达到了98%以上,准确地掌控了发酵时间。研究成果为面包机等烘焙设备的智能化发展提供了技术支持。
The fermentation process of dough is affected by many factors,such as the type of flour,temperature,amount of water added,amount of yeast,and is usually controlled by artificial experience and sensitivity.It is not only time-consuming and labourconsuming,but also has low accuracy.The machine vision technology was used,the image acquisition test bench was built,and the images of the dough fermentation process was captured. Five characteristics of the dough images were extracted and the dough fermentation time-domain process was determined:filling area-rising area-stagnation area. Decision tree algorithm was used to screen and train the data set,and the visualized tree of user-friendly and logically clear dough fermentation process was obtained.The 10-fold cross validation results show that the classification accuracy of dough fermentation process based on decision tree reached above 98%,thus the fermentation time was accurately mastered and controlled.The research results provide technical support for the intellectualization of bakery and other roasting equipments.
作者
孙美艳
刘峻
练毅
Sun Meiyan;Liu Jun;Lian Yi(Jiangsu University,Zhenjiang 212028,China;Jianghai Polytechnic College,Yangzhou 225101,China)
出处
《包装与食品机械》
CAS
北大核心
2019年第5期17-21,共5页
Packaging and Food Machinery
基金
江苏省产学研前瞻性联合研究项目（BY2016066-06）
江苏省普通高校学术学位研究生创新计划项目(KYLX16_0879)
关键词
面包机
面团发酵
决策树
机器视觉
智能化
数据挖掘
toaster
dough fermentation
decision tree
machine vision
intellectualization
data mining
