基于改进ResNet50模型的咖啡生豆质量和缺陷检测方法

doi:10.13733/j.jcam.issn.2095-5553.2024.04.034

摘要/Abstract

摘要： 咖啡生豆的质量决定着商品咖啡豆的价格，目前对咖啡生豆的筛选主要由人工完成，费时费力。提出一种基于改进ResNet50模型来识别咖啡生豆的方法，首先收集8 000张咖啡生豆图像建立数据集，并对其进行数据增强，基于ResNet50模型加入CBAM注意力机制，引入迁移学习机制，并使用深度可分离卷积来代替ResNet50残差单元中的传统卷积，构建适用于咖啡生豆分类识别的ResNet50CBAMDW模型。为评估模型改进的有效性，与ResNet50、AlexNet、VGG16、MobileNetV2等模型进行比较，改进后模型准确率达到91.1%，相较于原ResNet50模型准确率提升3.0%，参数量降低39.0%。

关键词: 残差网络, 咖啡豆, 注意力机制, 卷积神经网络, 深度可分离卷积

Abstract: The quality of raw coffee beans determines the price of commercial coffee beans. Currently, the screening of raw coffee beans is mainly done manually, which is timeconsuming and laborious. This paper proposes a method to identify raw coffee beans based on an improved ResNet50 model. Firstly, 8 000 images of raw coffee beans were collected to build a dataset and data enhancement was applied to it. A ResNet50CBAMDW model for coffee bean classification recognition was constructed based on the ResNet50 model by adding a CBAM attention mechanism, by introducing a migration learning mechanism and using deep separable convolution instead of the conventional convolution in the ResNet50 residual unit. In order to evaluate the effectiveness of the model improvement, the accuracy of the improved model was compared with ResNet50, AlexNet, VGG16, MobileNetV2 and other models, and the accuracy of the improved model reached 91.1%, which improved 3.0% compared with the original ResNet50 model and reduced the number of parameters by 39.0%.

Key words: residual networks, coffee beans, attention mechanism, convolutional neural networks, deep separable convolution

中图分类号:

纪元浩, 许金普, 严蓓蓓, 薛俊龙. 基于改进ResNet50模型的咖啡生豆质量和缺陷检测方法[J]. 中国农机化学报, 2024, 45(4): 237-243.

Ji Yuanhao, Xu Jinpu, Yan Beibei, Xue Junlong. A method for detecting quality and defects in raw coffee beans based on improved ResNet50 model[J]. Journal of Chinese Agricultural Mechanization, 2024, 45(4): 237-243.

参考文献

［1］　黄家雄, 黄琳, 吕玉兰, 等. 中国咖啡产业发展前景分析［J］. 云南农业科技, 2018(6): 4-7.
Huang Jiaxiong, Huang Lin, Lü Yulan, et al. Analysis of development prospect of coffee sector in China ［J］. Yunnan Agricultural Science and Technology, 2018(6): 4-7.
［2］　胡荣锁, 方乐天, 况沁蕊, 等. 云南临沧产区咖啡干香感官特征解析及杯品质量分析［J］. 食品科学, 2021, 42(20): 180-187.
Hu Rongsuo, Fang Letian, Kuang Qinrui, et al. Sensory characteristics of dry aroma and cupping quality of coffee from Lincang, Yunnan ［J］. Food Science, 2021, 42(20): 180-187.
［3］　Khuwijitjaru P. Near infrared spectroscopy research performance in food science and technology ［J］. NIR News, 2018, 29(3): 12-14.
［4］　Gomes W P C, Gonalves L, da Silva C B, et al. Application of multispectral imaging combined with machine learning models to discriminate special and traditional green coffee ［J］. Computers and Electronics in Agriculture, 2022, 198: 107097.
［5］　张珂, 冯晓晗, 郭玉荣, 等. 图像分类的深度卷积神经网络模型综述［J］. 中国图象图形学报, 2021, 26(10): 2305-2325.
Zhang Ke, Feng Xiaohan, Guo Yurong, et al. Overview of deep convolutional neural networks for image classification ［J］. Journal of Image and Graphics, 2021, 26(10): 2305-2325.
［6］　刘云玲, 张天雨, 姜明, 等. 基于机器视觉的葡萄品质无损检测方法研究进展［J］. 农业机械学报, 2022, 53(S1): 299-308.
Liu Yunling, Zhang Tianyu, Jiang Ming, et al. Review on nondestructive detection methods of grape quality based on machine vision ［J］. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(S1): 299-308.
［7］　赵辉, 乔艳军, 王红君, 等. 基于双通道注意力机制的ResNet果实外观品质分类［J］. 光电子·激光, 2022, 33(6): 643-651.
Zhao Hui, Qiao Yanjun, Wang Hongjun, et al. ResNet fruit appearance quality classification based on dual channel attention mechanism ［J］. Journal of Optoelectronics·Laser, 2022, 33(6): 643-651.
［8］　叶云, 赵小娟, 姜晟. 基于深度学习的荔枝虫害识别技术的应用与实现［J］. 中国农业信息, 2022, 34(4): 30-37.
Ye Yun, Zhao Xiaojuan, Jiang Sheng. Application and realization of litchi pest identification technology based on deep learning ［J］. China Agricultural Informatics, 2022, 34(4): 30-37.
［9］　Huang N F, Chou D L, Lee C A, et al. Smart agriculture: Realtime classification of green coffee beans by using a convolutional neural network ［J］. IET Smart Cities, 2020, 2(4): 167-172.
［10］　Adiwijaya N O, Romadhon H I, Putra J A, et al. The quality of coffee bean classification system based on color by using knearest neighbor method ［C］. Journal of Physics: Conference Series. IOP Publishing, 2022, 2157(1): 012034.
［11］　赵玉清, 杨慧丽, 张悦, 等. 基于特征组合与SVM的小粒种咖啡缺陷生豆检测［J］. 农业工程学报, 2022, 38(14): 295-302.
Zhao Yuqing, Yang Huili, Zhang Yue, et al. Detection of defective Arabica green coffee beans based on feature combination and SVM ［J］. Transactions of the Chinese Society of Agricultural Engineering, 2022, 38(14): 295-302.
［12］　Sandler M, Howard A, Zhu M, et al. Mobilenetv2: Inverted residuals and linear bottlenecks ［C］. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 4510-4520.
［13］　Febriana A, Muchtar K, Dawood R, et al. “USKCOFFEE Dataset: A multiclass green arabica coffee bean dataset for deep learning” ［C］. In 2022 IEEE International Conference on Cybernetics and Computational Intelligence (CyberneticsCom). IEEE, 2022.
［14］　郑远攀, 李广阳, 李晔. 深度学习在图像识别中的应用研究综述［J］. 计算机工程与应用, 2019, 55(12): 20-36.
Zheng Yuanpan, Li Guangyang, Li Ye. Survey of application of deep learning in image recognition ［J］. Computer Engineering and Applications, 2019, 55(12): 20-36.
［15］　He K, Zhang X, Ren S, et al. Deep residual learning for image recognition ［C］. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016: 770-778.
［16］　Woo S, Park J, Lee J Y, et al. CBAM: Convolutional block attention module ［C］. Proceedings of the European Conference on Computer Vision (ECCV), 2018: 3-19.
［17］　李书琴, 陈聪, 朱彤, 等. 基于轻量级残差网络的植物叶片病害识别［J］. 农业机械学报, 2022, 53(3): 243-250.
Li Shuqin, Chen Cong, Zhu Tong, et al. Plant leaf disease identification based on lightweight residual network ［J］. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(3): 243-250.
［18］　郑光, 魏家领, 任艳娜, 等. 基于深度可分离与空洞卷积的轻量化小麦生育进程监测模型研究［J］. 江苏农业科学, 2022, 50(20): 226-232.

[1]	霍正瑞, 孙铁波. 基于改进YOLOv5算法的瓜蒌分级方法[J]. 中国农机化学报, 2024, 45(4): 100-107.
[2]	林燕翔, 沈印, 李光林. 基于改进InceptionV3算法的小麦杂质识别研究[J]. 中国农机化学报, 2024, 45(4): 108-116.
[3]	陈鲁威, , 曾锦, , 袁全春, , 夏烨, , 潘健, 吕晓兰, , . 基于改进DeepLabV3+的梨树冠层分割方法[J]. 中国农机化学报, 2024, 45(4): 155-161.
[4]	严蓓蓓, 纪元浩, 曲凤凤, 许金普. 基于改进YOLOv5s的茶叶嫩芽检测[J]. 中国农机化学报, 2024, 45(4): 168-174.
[5]	张慧蒙, 何超, 徐嘉雯, 罗鑫, 荣剑, 刘学渊. 基于SCGYOLOv5n的收获期澳洲坚果检测算法[J]. 中国农机化学报, 2024, 45(4): 214-221.
[6]	邱金凯, 许秀英, 康烨, 臧浩, 马锴, 郭志鹏. 基于语义分割的无人机图像西瓜果实提取研究[J]. 中国农机化学报, 2024, 45(3): 182-188.
[7]	项新建, 郑雨, 曹光客, , 李旭, 尤钦寅, 姚佳娜. 基于改进Yolov5s的水稻叶病检测方法[J]. 中国农机化学报, 2024, 45(3): 212-218.
[8]	达措, , 赵启军, , , 高定国, , 索南尖措, 尼玛扎西. 基于注意力网络的长时牦牛个体识别研究[J]. 中国农机化学报, 2024, 45(1): 202-208.
[9]	黎远江, 李云伍, , 赵颖, , 台少瑜, 王克超. 基于改进Deeplabv3+模型的果树语义分割研究[J]. 中国农机化学报, 2024, 45(1): 209-216.
[10]	叶荣, 马自飞, 高泉, 李彤, 邵郭奇, 王白娟. 基于改进YOLOv5sECAASFF算法的茶叶病害目标检测[J]. 中国农机化学报, 2024, 45(1): 244-251.
[11]	王瑞彬, 杨世忠, 高升. 融合残差网络与注意力机制的草莓检测[J]. 中国农机化学报, 2024, 45(1): 266-273.
[12]	梁万杰, 曹静, 孙传亮, 曹宏鑫, 张文宇, . 基于深度学习的农作物病害识别系统研发[J]. 中国农机化学报, 2023, 44(9): 169-175.
[13]	杨波, 何金平, 张立娜. 基于改进SPP-x的YOLOv5神经网络水稻叶片病害识别检测[J]. 中国农机化学报, 2023, 44(9): 190-197.
[14]	汪健, 梁兴建, 雷刚. 基于深度残差网络与迁移学习的水稻病虫害图像识别[J]. 中国农机化学报, 2023, 44(9): 198-204.
[15]	余胜, 谢莉. 基于迁移学习和卷积视觉转换器的农作物病害识别研究[J]. 中国农机化学报, 2023, 44(8): 191-197.