食品与发酵工业 >

2021 , Vol. 47 >Issue 1: 259 - 265

DOI: https://doi.org/10.13995/j.cnki.11-1802/ts.024419

分析与检测

糖熏鸡腿颜色快速精准识别的多层卷积神经网络模型研究

  • 王博 ,
  • 杨洪遥 ,
  • 陆逢贵 ,
  • 陈子东 ,
  • 曹振霞 ,
  • 刘登勇
展开
  • 1(渤海大学 食品科学与工程学院,生鲜农产品贮藏加工及安全控制技术国家地方联合工程研究中心,辽宁 锦州,121013)
    2(哈尔滨商业大学 职业技术教育学院,黑龙江 哈尔滨,150028)
    3(江苏省肉类生产与加工质量安全控制协同创新中心,江苏 南京,210095)
王博讲师和杨洪遥硕士研究生为共同第一作者(刘登勇教授为通讯作者,E-mail: jz_dyliu@126.com)

收稿日期: 2020-05-10

  修回日期: 2020-08-10

  网络出版日期: 2021-02-03

基金资助

国家重点研发计划项目(2016YFD0401505);辽宁省“兴辽英才计划”项目(XLYC1 807 100);辽宁省一流学科项目(LNSPXKBD2020204,LNSPXKBD2020306)

收起

Vision research on CNN model for quick and accurate identification of sugar-smoked chicken thighs color

  • WANG Bo ,
  • YANG Hongyao ,
  • LU Fenggui ,
  • CHEN Zidong ,
  • CAO Zhenxia ,
  • LIU Dengyong
Expand
  • 1(National & Local Joint Engineering Research Center of Storage, Processing and Safety Control Technology for Fresh Agricultural and Aquatic Products, College of Food Science and Technology, Bohai University, Jinzhou 121013, China)
    2(School of Vocational and Technical Education, Harbin University of Commerce, Harbin 150028, China)
    3(Jiangsu Collaborative Innovation Center of Meat Production and Processing, Quality and Safety Control, Nanjing 210095, China)

Received date: 2020-05-10

  Revised date: 2020-08-10

  Online published: 2021-02-03

Fold

摘要

为快速精准识别糖熏鸡腿在熏制过程中产生的所有颜色,基于机器视觉技术,构建Xception-CNN模型用于熏鸡腿颜色的识别,同时应用ResNet-50、Inception和传统卷积神经网络(convolutional neural networks,CNN)等3种模型对比分析Xception-CNN模型对熏鸡腿颜色的识别效果。采集并经过图像预处理后,共得到不同颜色的熏鸡腿图像4 352张,作为4种模型的实验样本,随机选取其中的3 482张作为训练组,剩下的870张作为测试组。结果表明,4种模型的平均识别准确率分别为92%(Xception-CNN)、91%(ResNet-50)、89%(Inception)、87%(传统CNN);测试时间分别为1.36 s(Xception-CNN)、0.81 s(ResNet-50)、0.98 s(Inception)、2.48 s(传统CNN)。Xception-CNN模型对糖熏鸡腿图像的颜色识别准确率最高,达到92%,测试时间略高于ResNet-50模型和Inception模型,但低于传统CNN模型,仅需1.36 s即可完成识别,此模型可以实现糖熏鸡腿颜色的快速精准识别,为糖熏工艺参数精准调控、保障产品颜色标准化等提供可靠依据。

关键词： 熏鸡; 糖熏; 颜色识别; 机器视觉; 多层卷积神经网络模型

本文引用格式

王博 , 杨洪遥 , 陆逢贵 , 陈子东 , 曹振霞 , 刘登勇 . 糖熏鸡腿颜色快速精准识别的多层卷积神经网络模型研究[J]. 食品与发酵工业, 2021 , 47(1) : 259 -265 . DOI: 10.13995/j.cnki.11-1802/ts.024419

Abstract

To explore the machine vision model that can quickly and accurately identify all the colors of smoked chicken legs during the smoking process, this article builds an Xception-CNN model for the recognition of smoked chicken leg colors based on machine vision technology, while compares and analyzes the Xception-CNN model’s recognition effect on smoked chicken leg color by applying ResNet-50, Inception and traditional CNN and other three models.A total of 4 352 smoked chicken leg images of different colors were collected and image pre-processed as experimental samples of four models, of which 3 482 were randomly selected as the training group, and the remaining 870 as the test group.The average recognition accuracy of the four models is Xception-CNN-92%, ResNet-50-91%, Inception-89%, traditional CNN-87%, and test time is Xception-CNN-1.36 s, ResNet-50-0.81 s, Inception-0.98 s, traditional CNN-2.48 s.Therefore, the Xception-CNN model has the highest color recognition accuracy for sugar-smoked chicken leg images, reaching 92%.The test time is slightly higher than the ResNet-50 model and the Inception model, but lower than the traditional CNN model.It only takes 1.36 s to complete the recognition.The fast and accurate identification of the color of smoked chicken legs can provide a reliable basis for the precise control of the parameters of the sugar smoked process and the standardization of product colors.

Key words： smoked chicken; sugar smoked; color recognition; machine vision; multilayer convolutional neural network model

参考文献

[1] 张顺亮,王守伟,成晓瑜,等.湖南腊肉加工过程中挥发性风味成分的变化分析[J].食品科学,2015,36(16):215-219.
ZHANG S L,WANG S W,CHENG X Y,et al.Changes in volatile flavor components during hunan cured meat processing[J].Food Science,2015,36(16):215-219.
[2] HITZEL A,MARGARETE P,FREDI S,et al.Polycyclic aromatic hydrocarbons (PAH) and phenolic substances in meat products smoked with different types of wood and smoking spices[J].Food Chemistry,2013,139(1-4):955-962.
[3] KERRY J P,LEDWARD D,KERRY J P,et al.Index-Improving the sensory and nutritional quality of fresh meat[M].London:Woodhead Publishing and CRC Press,2009.
[4] MANCINI R A,HUNT M C.Current research in meat color[J].Meat Science,2005,71(1):100-121.
[5] ZHUANG H,BOWKER B.Effect of marination on lightness of broiler breast fillets varies with raw meat color attributes[J].LWT-Food Science and Technology,2016,69:233-235.
[6] 刘登勇, 吴金城,王继业,等.沟帮子熏鸡腿主体风味成分分析[J].食品工业科技,2018,39(7):237-242.
LIU D Y,WU J C,WANG J Y,et al.Analysis of key odor compounds of Goubangzi smoked chicken[J].Science and Technology of Food Industry,2018,39(7):237-242.
[7] 段艳, 郑福平,杨梦云,等.ASE-SAFE/GC-MS/GC-O法分析德州扒鸡风味化合物[J].中国食品学报,2014,14(4):222-230.
DUAN Y,ZHENG F P,YANG M Y,et al.Analysis on volatile flavor compounds in Dezhou braised chicken by ASE-SAFE/GC-MS/GC-O[J].Journal of Chinese Institute of Food Science and Technology,2014,14(4):222-230.
[8] CHEN S,KAO T H,CHEN C J,et al.Reduction of carcinogenic polycyclic aromatic hydrocarbons in meat by sugar-smoking and dietary exposure assessment in Taiwan[J].Journal of Agricultural and Food Chemistry,2013,61(31):7 645-7 653.
[9] 姜沛宏, 张玉华,钱乃余,等.基于机器视觉技术的肉新鲜度分级方法研究[J].食品科技,2015,40(3):296-300.
JIANG F H,ZHANG Y H,QIAN N Y,et al.Research on method to freshness grading of meat based on machine vision technology[J].Food Science and Technology,2015,40(3):296-300.
[10] 王树才,陶凯,李航.基于机器视觉定位的家禽屠宰净膛系统设计与试验[J].农业机械学报,2018,49(1):335-343.
WANG S C,TAO K,LI H.Design and experiment of poultry eviscerator system based on machine vision positioning[J].Transactions of the Chinese Society for Agricultural Machinery,2018,49(1):335-343.
[11] TU K L,LI L J,YANG L M,et al.Selection for high quality pepper seeds by machine vision and classifiers[J].Journal of Integrative Agriculture,2018,17(9):1 999-2 006.
[12] 李文采, 李家鹏,田寒友,等.基于RGB颜色空间的冷冻猪肉储藏时间机器视觉判定[J].农业工程学报,2019,35(3):294-300.
LI W C,LI J P,TIAN H Y,et al.Determination of storage time for chilled pork by using RGB color space method based on machine vision[J].Transactions of the Chinese Society of Agricultural Engineering,2019,35(3):294-300.
[13] SEO J K,BAIK S B,LEE S H,et al.Simulating the architecture of a termite incipient nest using a convolutional neural network[J].Ecological Informatics,2018(44):94-100.
[14] SHEN Y F,ZHOU H L,LI J T,et al.Detection of stored-grain insects using deep learning[J].Computers and Electronics in Agriculture,2018,145:319-325.
[15] XU J L,SUN D W.Computer vision detection of salmon muscle gaping using convolutional neural network features[J].Food Analytical Methods,2018,11(1):34-47.
[16] PARITOSH P,D AKELLA,M BAPPADITYA,et al.FoodNet:Recognizing foods using ensemble of deep networks[J].IEEE Signal Processing Letters,2017,24(12):1 758-1 762.
[17] FU Z H,CHEN D,LI H Y.A large benchmark dataset for chinese food recognition [C].International Conference on Intelligent Computing,2017(10361):273-281.
[18] 梅舒欢, 闵巍庆,刘林虎,等.基于Faster R-CNN的食品图像检索和分类[J].南京信息工程大学学报(自然科学版),2017,9(6):635-641.
MEI S H,MIN W Q,LIU L H,et al.Faster R-CNN based food image retrieval and classification[J].Journal of Nanjing University of Information Science & Technology(Natural Science Edition), 2017,9(6):635-641.
[19] AMANDA R,KELSEE B,PETER M,et al.Deep learning for image-based cassava disease detection[J].Frontiers In Plant Science,2017(8):1 852.
[20] MCALLISTER P,ZHENG HR,R BOND,et al.Combining deep residual neural network features with supervised machine learning algorithms to classify diverse food image datasets[J].Computers In Biology And Medicine,2018(95):217-233.
[21] BAI L,ZHAO Y M,HUANG X M.A CNN accelerator on FPGA using depthwise separable convolution[J].IEEE Transactions On Circuits And Systems II-Express Briefs,2018,65(10):1 415-1 419.
[22] CHOLLET F.Xception:Deep learning with depthwise separable convolutions[C].2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).IEEE,2017:1610.02357.
[23] KONOVALOV D A,SALEH A,BRADLEY M,et al.Underwater fish detection with weak multi-domain supervision[C].2019 International Joint Conference on Neural Networks (IJCNN).2019.
[24] 黄现青, 董飒爽,李传令,等.冷却鸡胸肉脉冲强光杀菌参数试验优化[J].农业机械学报,2019,50(2):333-339.
HUANG X Q,DONG S S,LI C L,et al.Effects of pulsed light parameters on bacterium sterilization and quality of chilled chicken breast meat[J].Transactions of the Chinese Society for Agricultural Machinery,2019,50(2):333-339.
[25] 彭增起, 徐幸莲,周光宏,等.一种猪肉颜色质量等级的划分方法:中国,CN1751812 [P].2006-03-29.
PENG Z Q,XU X L,ZHOU G H,et al.Method for grading pork based on its color and quality:China,CN1751812 [P].2006-03-29.
[26] GHASEMI-VARNAMKHASTI M,GOLI R,FORINA M,et al.Application of image analysis combined with computational expert approaches for shrimp freshness evaluation[J].International Journal of Food Properties,2016,19(10):2 202-2 222.
[27] LECUN Y,BOSER B,DENKER J S,et al.Backpropagation applied to handwritten zip code recognition[J].Neural computation,1989,1(4):541-551.
[28] 王巧华, 王彩云,马美湖.基于机器视觉的鸭蛋新鲜度检测[J].中国食品学报,2017,17(8):268-274.
WANF Q H,WANG C Y,MA M H.Duck eggs’freshness detection based on machine vision technology[J].Journal of Chinese Institute of Food Science and Technology,2017,17(8):268-274.
[29] SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[C].Published as a conference paper at ICLR,2015,2015:1-14.
[30] 黄旭, 凌志刚,李绣心.融合判别式深度特征学习的图像识别算法[J].中国图象图形学报,2018,23(4):510-518.
HUAGN X,LING Z G,LI X X.Discriminative deep feature learning method by fusing linear discriminant analysis for image recognition[J].Journal of Image and Graphics,2018,23(4):510-518.
Options
文章导航

/

〈 〉

技术支持：北京玛格泰克科技发展有限公司