Abstract
Objective: To solve the problems of poor recognition accuracy and low efficiency of existing automatic recognition methods for packaged food. Methods: Based on the automatic identification system of packaged food, an improved YOLOv3 model was proposed for the automatic identification of soft packaged food. The Kmeans++algorithm was introduced into the model to solve the problem of small target insensitivity, the Mish activation function was introduced into the model to improve the accuracy of recognition, and the attention mechanism Senet was introduced into the model to improve the ability of feature extraction. The performance of the recognition model was analyzed through experiments, and the superiority of the model was verified. Results: Compared with the conventional recognition methods, the proposed method can more accurately and efficiently realize the automatic recognition of flexible packaging food, the recognition accuracy rate was 95.40%, and the recognition efficiency was 23.80 f/s, which meets the needs of packaging food recognition. Conclusion: By optimizing the existing food recognition model, the performance of the recognition model can be effectively improved.
Publication Date
6-9-2023
First Page
95
Last Page
100
DOI
10.13652/j.spjx.1003.5788.2022.60167
Recommended Citation
Zhi-kai, ZHANG; Hong-zhang, HAN; Xue-qian, ZHAO; and Zhong, LI
(2023)
"Automatic recognition method for soft packaged food based on improved YOLOv3 model,"
Food and Machinery: Vol. 39:
Iss.
5, Article 16.
DOI: 10.13652/j.spjx.1003.5788.2022.60167
Available at:
https://www.ifoodmm.cn/journal/vol39/iss5/16
References
[1] 郝琳, 张坤平. 基于并联机器人的食品分拣控制系统设计[J]. 食品工业, 2020, 41（4）: 209-212.
HAO L, ZHANG K P. Design of food sorting control system based on parallel robot[J]. The Food Industry, 2020, 41（4）: 209-212.
[2] 杨志锐, 郑宏, 郭中原, 等. 基于网中网卷积神经网络的红枣缺陷检测[J]. 食品与机械, 2020, 36（2）: 140-145, 181.
YNAG Z R, ZHENG H, GUO Z Y, et al. Chinese date defect detection based on net in net convolution neural network[J]. Food & Machinery, 2020, 36（2）: 140-145, 181.
[3] 赵腾飞, 胡国玉, 周建平, 等. 卷积神经网络算法在核桃仁分类中的研究[J]. 中国农机化学报, 2022, 43（6）: 181-189.
ZHAO T F, HU G Y, ZHOU J P, et al. Research on convolution neural network algorithm in walnut kernel classification[J]. Journal of Chinese Agricultural Mechanization, 2022, 43（6）: 181-189.
[4] 吕铷麟, 贾镇, 胡益滔, 等. 基于卷积神经网络的食品塑料包装袋光谱识别[J]. 包装工程, 2022, 43（3）: 121-128.
LU R B, JIA Z, HU Y T, et al. Spectral recognition of food plastic packaging bags based on convolution neural network[J]. Packaging Engineering, 2022, 43（3）: 121-128.
[5] 郑如新, 孙青云, 马素慧, 等. 基于高斯混合模型的水果分类识别方法研究[J]. 制造业自动化, 2022, 44（12）: 16-19.
ZHENG R X, SUN Q Y, MA S H, et al. Research on fruit classification and recognition method based on Gaussian mixture model[J]. Manufacturing Automation, 2022, 44（12）: 16-19.
[6] 刘宇昕, 闵巍庆, 蒋树强, 等. 多尺度拼图重构网络的食品图像识别[J]. 软件学报, 2022, 33（11）: 4 379-4 395.
LIU Y X, MIN W Q, JIANG S Q, et al. Food image recognition based on multi-scale mosaic reconstruction network[J]. Journal of Software, 2022, 33（11）: 4 379-4 395.
[7] 严培培. 面向非典型食品生产的高速机器人分拣系统设计[J]. 食品与机械, 2016, 32（2）: 94-97.
YAN P P. Design of high-speed robot sorting system for atypical food production[J]. Food & Machinery, 2016, 32（2）: 94-97.
[8] 柳振宇, 薛毓强, 谢祖强. 基于闭环和前馈控制的高速食品分拣机器人控制技术[J]. 食品与机械, 2021, 37（7）: 87-93.
LIU Z Y, XUE Y Q, XIE Z Q. Control technology of high-speed food sorting robot based on closed-loop and feedforward control[J]. Food & Machinery, 2021, 37（7）: 87-93.
[9] 吴旭清, 黄家才, 周磊, 等. 并联机器人智能分拣系统设计[J]. 机电工程, 2019, 36（2）: 224-228.
WU X Q, HUANG J C, ZHOU L, et al. Design of intelligent sorting system for parallel robot[J]. Electromechanical Engineering, 2019, 36（2）: 224-228.
[10] 倪鹤鹏, 刘亚男, 张承瑞, 等. 基于机器视觉的Delta机器人分拣系统算法[J]. 机器人, 2016, 38（1）: 49-55.
NI H P, LIU Y N, ZHANG C R, et al. Algorithm of delta robot sorting system based on machine vision[J]. Robot, 2016, 38（1）: 49-55.
[11] 郝大孝, 舒志兵, 孙学. 基于机器视觉的 Delta 机器人分拣与跟踪系统设计[J]. 机床与液压, 2019, 47（17）: 36-42.
HAO D X, SHU Z B, SUN X. Design of Delta robot sorting and tracking system based on machine vision[J]. Machine Tool and Hydraulic, 2019, 47（17）: 36-42.
[12] 刘芳, 刘玉坤, 林森, 等. 基于改进型YOLO的复杂环境下番茄果实快速识别方法[J]. 农业机械学报, 2020, 51（6）: 229-237.
LIU F, LIU Y K, LIN S, et al. Rapid identification method of tomato fruit in complex environment based on improved Yolo[J]. Acta Agriculturalis Sinica, 2020, 51（6）: 229-237.
[13] 赵利平, 吴德刚. 融合GA的三点定位夜间苹果目标的识别算法研究[J]. 中国农机化学报, 2020, 41（5）: 134-138.
ZHAO L P, WU D G. Research on recognition algorithm of three-point positioning night apple target based on GA[J]. China Agricultural Machinery Chemical Journal, 2020, 41（5）: 134-138.
[14] 朱建宝, 许志龙, 孙玉玮, 等. 基于 OpenPose 人体姿态识别的变电站危险行为检测[J]. 自动化与仪表, 2020, 35（2）: 47-51.
ZHU J B, XU Z L, SUN Y W, et al. Detection of dangerous behaviors in substations based on OpenPose human posture recognition[J]. Automation and Instrumentation, 2020, 35（2）: 47-51.
[15] 陶浩, 李笑, 陈敏. 基于改进ORB特征的遥操作工程机器人双目视觉定位[J]. 测控技术, 2019, 38（7）: 19-23.
TAO H, LI X, CHEN M. Binocular vision of teleoperation engineering robot based on improved ORB feature[J]. Measurement and Control Technology, 2019, 38（7）: 19-23.
[16] 宋海涛, 何文浩, 原魁. 一种基于SIFT特征的机器人环境感知双目立体视觉系统[J]. 控制与决策, 2019, 34（7）: 1 545-1 552.
SONG H T, HE W H, YUAN K. A robot environment perception binocular stereo vision system based on SIFT feature[J]. Control and Decision, 2019, 34（7）: 1 545-1 552.
[17] 马伟苹, 李文新, 孙晋川, 等. 基于粗—精立体匹配的双目视觉目标定位方法[J]. 计算机应用, 2020, 40（1）: 227-232.
MA W P, LI W X, SUN J C, et al. Binocular vision target location method based on coarse fine stereo matching[J]. Computer Applications, 2020, 40（1）: 227-232.
[18] 罗久飞, 邱广, 张毅, 等. 基于自适应双阈值的SURF双目视觉匹配算法研究[J]. 仪器仪表学报, 2020, 41（3）: 240-247.
LUO J F, QIU G, ZHANG Y, et al. Research on surf binocular vision matching algorithm based on adaptive double threshold[J]. Journal of Instrumentation, 2020, 41（3）: 240-247.
[19] 程禹, 王晓华, 王文杰, 等. 基于改进 AKAZE算法的图像特征匹配方法[J]. 西安工程大学学报, 2020, 34（8）: 51-56.
CHENG Y, WANG X H, WANG W J, et al. Image feature matching method based on improved AKAZE algorithm[J]. Journal of Xi'an Engineering University, 2020, 34（8）: 51-56.
[20] 曾劲松, 薛文凯, 徐博凡, 等. 双目视觉引导机器人定位抓取技术的研究[J]. 组合机床与自动化加工技术, 2019, 12（1）: 131-137.
ZENG J S, XUE W K, XU B F, et al. Research on positioning and grasping technology of binocular vision guided robot[J]. Modular Machine Tools and Automatic Machining Technology, 2019, 12（1）: 131-137.
[21] 林义忠, 陈旭. 基于机器视觉的机器人定位抓取的研究进展[J]. 自动化与仪器仪表, 2021, 11（3）: 9-12.
LIN Y Z, CHEN X. Research progress of robot positioning and grasping based on machine vision[J]. Automation and Instrumentation, 2021, 11（3）: 9-12.
[22] 蒋镕圻, 彭月平, 谢文宣, 等. 嵌入scSE模块的改进 YOLOv4 小目标检测算法[J]. 图学学报, 2021, 42（4）: 546-555.
JIANG R Q, PENG Y P, XIE W X, et al. Improved YOLOv4 small target detection algorithm embedded in scSE module[J]. Journal of Graphics, 2021, 42（4）: 546-555.