基于BP神经网络的面包工业化生产发酵程度视觉判定  

Study on visual judgment method of fermentation degree in industrialized bread production based on BP neural network

作　　者：赵纯 陈学永[1] 吴少霜 庞杰[2] ZHAO Chun;CHEN Xue-yong;WU Shao-shuang;PANG Jie(College of Mechanical and Electrical Engineering,Fujian Agriculture and Forestry University,Fuzhou,Fujian 350000,China;College of Food Science,Fujian Agriculture and Forestry University,Fuzhou,Fujian 350000,China)

机构地区：[1]福建农林大学机电工程学院,福建福州350000 [2]福建农林大学食品科学学院,福建福州350000

出　　处：《食品与机械》2022年第2期197-202,共6页Food and Machinery

基　　金：国家自然科学基金(编号:31772045)。

摘　　要：目的:建立BP神经网络模型来实现对面团发酵成熟度的精准判定。方法:利用机器视觉技术采集并处理面包发酵过程的图片,获取图片的量化信息,建立了一种基于BP(back propagation)神经网络的面包发酵程度预测模型。将时间、面积、瞬时速度、膨胀率、灰度值能量、灰度值相互关系、灰度值均匀性、灰度值对比度,作为输入神经元,建立相关参数的回归模型。结果:给定特征参数与面包的发酵过程存在相关性,能够有效表征面包的发酵程度;基于BP神经网络建立的预测模型准确率达到88.41%。结论:该模型能够实现对面包发酵程度的准确判定,有效提高面包工业化生产中品质的一致性。Objective: A BP neural network model was established to accurately determine the maturity of dough fermentation. Methods: Used machine vision technology to collect and process the pictures of bread fermentation process, obtained the quantitative information of the pictures, and a prediction model was established for bread fermentation degree based on BP neural network. Taking the time, area, instantaneous velocity, expansion rate, gray value energy, gray value relationship, gray value uniformity and gray value contrast as the input neurons, the regression model of relevant parameters was established. Results: The correlation between the above characteristic parameters and the fermentation process of bread was determined, which can effectively characterize the fermentation degree of bread;The results show that the accuracy of the prediction model based on BP neural network is 88.41%. Conclusion: The model can accurately determine the degree of bread fermentation and effectively improve the quality consistency in industrial bread production.

关 键 词：机器视觉 BP神经网络 发酵 面包 计算机视觉 

分 类 号：TS213.21[轻工技术与工程—粮食、油脂及植物蛋白工程] TP183[轻工技术与工程—食品科学与工程]