基于改进ResNet的带孔蛋胚裂纹检测方法#br#

doi:10.13733/j.jcam.issn.2095-5553.2024.07.030

摘要/Abstract

摘要： 将病毒注入蛋胚培养疫苗时，针头冲击使注射孔周围产生裂纹导致培养失败。为解决目前人工手持照蛋器在暗室通过肉眼检测蛋胚裂纹效率低、误判率高这一问题，提出一种结合多光谱通道注意(MSCA)机制与并行堆叠拓扑(PST)模块的带孔蛋胚裂纹检测方法。首先搭建带孔蛋胚裂纹检测黑箱，采集病毒注射后带孔蛋胚的线形、网状、星形裂纹和带孔完好蛋胚图像并建立数据集；接着以ResNet-50为骨干模型，将其后4层的第一个残差模块替换为PST模块，以增加模型初期图像表达能力；最后在每个PST模块与残差模块后引入MSCA机制，MSCA机制通过二维离散余弦变换(2DDCT)压缩数据得到各通道频率分量，采用神经结构搜索(NAS)方式得到最佳频率分量，即对权重重新分配，增加裂纹特征权重比例，以确保模型快速、精准识别带孔蛋胚裂纹。结果表明，改进的网络模型对带孔蛋胚裂纹检测时间为0.42 s/枚，检测精度为96.43%，检测效率高于人工作业。与原始ResNet-50相比，检测精度提高3.66%，优于其他经典卷积网络模型，改进模型可用于疫苗培育前带孔蛋胚裂纹自动检测。

关键词: 裂纹检测, 带孔蛋胚, 改进残差网络, 多光谱通道注意机制, 并行堆叠拓扑模块

Abstract: In the process of injecting the virus into the egg embryo to culture the vaccine, the impact of the needle caused cracks around the injection hole and the culture failed. In order to solve the problem of low efficiency and high misjudgment rate in the detection of egg embryo cracks by the naked eye in the darkroom, a method for detecting cracks in egg embryos with holes combined with a multispectral channel attention mechanism (MSCA) and parallel stacking topology modules (PST) was proposed. Firstly, a black box for crack detection of perforated egg embryos was built, and the images of linear cracks, mesh cracks, starshaped cracks and intact egg embryos with holes after virus injection was collected and a data set was established, then ResNet-50 was used as the backbone model. The first residual module of the next 4 layers was replaced by the PST module to increase the image expression ability of the model at the beginning, finally, the MSCA mechanism was introduced after each PST module and the residual module, and the MSCA mechanism compressed the data through twodimensional discrete cosine transform (2D DCT), and the neural architecture search selection (NAS) method obtained the optimal frequency components, which could redistribute the weights to increase the proportion of crack feature weights. Thus the microporous egg embryos could be identified quickly and accurately by the model. The results show that the improved network model has a detection time of 0.42 s/piece for cracks in egg embryos with holes, a detection accuracy of 96.43%, and a higher detection efficiency than manual operations. Compared with the original ResNet-50, the detection accuracy has been improved by 3.66%, which is superior to other classical convolutional network models. It is proved that the improved model can be used for the automatic detection of cracks in egg embryos with holes before vaccine cultivation.

Key words: crack detection, egg embryo with hole, improved ResNet, multispectral attention mechanism, parallel stacking topology

中图分类号:

TP391.4

李运良, 赵明岩, 王鑫, 闫泓硕, 李育蝉. 基于改进ResNet的带孔蛋胚裂纹检测方法#br#[J]. 中国农机化学报, 2024, 45(7): 201-208.

Li Yunliang, Zhao Mingyan, Wang Xin, Yan Hongshuo, Li Yuchan. Detection method of egg embryo crack with hole based on improved ResNet[J]. Journal of Chinese Agricultural Mechanization, 2024, 45(7): 201-208.

参考文献

［1］　Iseki H, Watanabe S, Mase M. A potential system for the isolation and propagation of porcine deltacoronavirus using embryonated chicken eggs ［J］. Journal of Virological Methods, 2021, 290: 114068.
［2］　Yu Qingzhong, Li Yufeng, Kiril Dimitrov, et al. Genetic stability of a Newcastle disease virus vectored infectious laryngotracheitis virus vaccine after serial passages in chicken embryos ［J］. Vaccine, 2019, 38(4): 925-932.
［3］　唐静, 辛丽, 李晓丹, 等. 2015—2016年度中国甲型H1N1亚型流感病毒疫苗候选株的制备及鉴定［J］. 中华实验和临床病毒学杂志, 2017, 31(4): 348-352.Tang Jing, Xin Li, Li Xiaodan, et al. Preparation and identification of influenza H1N1 subtype vaccine candidate strain in China ［J］. Chinese Journal of Experimental and Clinical Virology, 2017, 31(4): 348-352.
［4］　徐彦伟, 徐爱军, 颉潭成, 等. 基于多信息融合的疫苗制备中鸡蛋胚体分拣系统［J］. 农业机械学报, 2015, 46(2): 20-26.
Xu Yanwei, Xu Aijun, Xie Tancheng, et al. Automatic sorting system of egg embryo in biological vaccines production based on multiinformation fusion ［J］. Transactions of the Chinese Society for Agricultural Machinery, 2015, 46(2): 20-26.
［5］　郭盟, 董新明, 韩广, 等. 基于卷积神经网络的蛋胚活性精准检测方法研究［J］. 应用光学, 2021, 42(2): 268-275.Guo Meng, Dong Xinming, Han Guang, et al. Research on accurate detection method of egg embryo activity based on convolutional neural network ［J］. Journal of Applied Optics, 2021, 42(2): 268-275.
［6］　Bao G J, Jia M M, Xun Y, et al. Cracked egg recognition based on machine vision ［J］. Computers and Electronics in Agriculture, 2019, 158: 159-166.
［7］　涂伟沪, 蔡玲霞, 李学军. 基于改进蝗虫算法优化Canny算子的鸡蛋裂纹图像检测［J］. 食品与机械, 2022, 38(2): 167-172, 202.
Tu Weihu, Cai Lingxia, Li Xuejun. Egg crack image detection method based on improved grasshopper optimization algorithm and Canny operator ［J］. Food & Machinery, 2022, 38(2): 167-172, 202.
［8］　辛永信, 黄泳波. 基于机器视觉的禽蛋脏污及裂纹检测系统设计［J］. 机电信息, 2020(2): 108-110.
［9］　Cheng C W, Feng P H, Xie J H, et al. Eggshell crack detection and egg classification using resonance and support vector machine methods ［J］. Applied Engineering in Agriculture, 2019, 35(1): 23-30.
［10］　Lai C C, Li C H, Huang K J, et al. Duck eggshell crack detection by nondestructive sonic measurement and analysis ［J］. Sensors, 2021, 21(21): 7299.
［11］　刘鹏, 屠康, 潘磊庆, 等. 基于多传感器融合的鸡蛋裂纹系统性识别［J］. 农业机械学报, 2010, 41(10): 185-189, 203.

Liu Peng, Tu Kang, Pan Leiqing, et al. Systematic recognition research of egg crack based on multisensor fusion ［J］. Transactions of the Chinese Society for Agricultural Machinery, 2010, 41(10): 185-189, 203.

［12］　Sun Li, Zhang Pengqi, Feng Siyu, et al. Eggshell crack detection based on the transient impact analysis and crosscorrelation method ［J］. Current Research in Food Science, 2021, 4: 716-723.
［13］　罗慧, 闫思蒙, 卢伟, 等. 基于力—声学特性的鸡蛋带小裂纹在线检测方法［J］. 农业机械学报, 2016, 47(11): 224-230.
Luo Hui, Yan Simeng, Lu Wei, et al. Microcracked eggs online detection method based on forceacoustic features ［J］. Transactions of the Chinese Society for Agricultural Machinery, 2016, 47(11): 224-230.
［14］　李雅琪, 孙力, 陈诚, 等. 基于多维振动响应信号分析的鸡蛋裂纹检测研究［J］. 镇江高专学报, 2019, 32(4): 31-35.Li Yaqi, Sun Li, Chen Cheng, et al. A study of eggshell crack detection based on the analysis of multidimensional vibration signals ［J］. Journal of Zhenjiang College, 2019, 32(4): 31-35.
［15］　陈静, 梁俊毅. 基于机器视觉的禽蛋外观品质在线检测系统设计［J］. 食品工业, 2020, 41(5): 232-234.Chen Jing, Liang Junyi. Design of online detection system for egg appearance quality based on machine vision ［J］. The Food Industry, 2020, 41(5): 232-234.
［16］　赵祚喜, 罗阳帆, 黄杏彪, 等. 基于机器视觉和YOLOv4的破损鸡蛋在线检测研究［J］. 现代农业装备, 2022, 43(1): 8-16.Zhao Zuoxi, Luo Yangfan, Huang Xingbiao, et al. Research on online detection of damaged eggs based on machine vision and YOLOv4 ［J］. Modern Agricultural Equipment, 2022, 43(1): 8-16.
［17］　张健, 崔英杰. 基于改进粒子群算法的鸡蛋裂纹检测方法［J］. 食品与机械, 2020, 36(7): 136-139, 226.〖JP3〗Zhang Jian, Cui Yingjie. Egg crack detection based on improved particle swarm optimization ［J］. Food & Machinery, 2020, 36(7): 136-139, 226.
［18］　He K M, Zhang X Y, Ren S Q, et al. Deep residual learning for image recognition ［C］. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 770-778.
［19］　Guo M H, Xu T X, Liu J J, et al. Attention mechanisms in computer vision: A survey ［J］. Computational Visual Media, 2022: 1-38.
［20］　Qin Z Q, Zhang P Y, Wu F, et al. FcaNet: Frequency channel attention networks ［C］. Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021: 783-792.
［21］　EscuderoVinolo M, Bescos J. Squeezing the DCT to fight camouflage ［J］. Journal of Mathematical Imaging and Vision, 2020, 62(2): 206-222.
［22］　万丰, 苑豪杰, 宫威. 基于离散余弦变换的数字图像压缩技术研究［J］. 自动化应用, 2020(3): 65-67.
［23］　狄恩彪, 徐光辉. 基于部分离散余弦变换的卷积神经网络设计与分析［J］. 通信技术, 2020, 53(7): 1636-1640.
Di Enbiao, Xu Guanghui. Design and analysis of convolutional neural network based on partial discrete cosine transform ［J］. Communications Technology, 2020, 53(7): 1636-1640.
［24］　Xie S N, Girshick R, Dollar P, et al. Aggregated residual transformations for deep neural networks ［C］. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 1492-1500.

[1]	王雷, 钟康生, 郭小宝, 盛斌, 肖波, 荆磊. 基于机器视觉的陈皮热泵干燥含水率预测模型研究[J]. 中国农机化学报, 2024, 45(7): 97-103.
[2]	程麒文, 邱白晶. 灰飞虱自动识别计数方法研究[J]. 中国农机化学报, 2024, 45(7): 166-171.
[3]	黄晓宇, 张聪, 陈晓玲. 融合多尺度特征的遮挡番茄病害图像识别研究[J]. 中国农机化学报, 2024, 45(7): 194-200.
[4]	黄亦其, 鹿林飞, 沈豪, 王福宽, 乔曦, . 基于卷积神经网络的入侵昆虫识别研究[J]. 中国农机化学报, 2024, 45(7): 222-227.
[5]	张晨一, 张晓乾, 任振辉. 基于改进YOLOv3的自然场景中未成熟青苹果识别研究[J]. 中国农机化学报, 2024, 45(7): 243-248.
[6]	熊志刚, 陈为真. 一种YOLOv5-DK的柠檬初期病虫害检测算法研究[J]. 中国农机化学报, 2024, 45(7): 249-254.
[7]	张书贵, 陈书理, 赵展. 改进YOLOv8的农作物叶片病虫害识别算法[J]. 中国农机化学报, 2024, 45(7): 255-260.
[8]	张继成, 黄向党. 基于改进YOLOv3的玉米病害识别方法[J]. 中国农机化学报, 2024, 45(7): 269-275.
[9]	侯海敏, 沈梦姣, 张艳. 作物病害检测关键技术分析[J]. 中国农机化学报, 2024, 45(6): 90-97.
[10]	雷浩, 苑迎春, 何振学. 基于多尺度卷积和注意力机制的枣品种识别[J]. 中国农机化学报, 2024, 45(6): 135-141.
[11]	张煜恒, 周宏平. 基于激光点云的树木特征信息提取研究进展[J]. 中国农机化学报, 2024, 45(6): 142-148.
[12]	杨昊霖, 王其欢, 李华彪, 耿端阳, 武继达, 姚艳春. 基于改进YOLOv5的田间复杂环境障碍物检测[J]. 中国农机化学报, 2024, 45(6): 216-222.
[13]	黄尧, 何敬, 付饶, 刘刚, , 林远杨. YOLOv5s-CBAM算法在福寿螺虫卵识别中的应用分析[J]. 中国农机化学报, 2024, 45(6): 223-228.
[14]	石放, 王莹, 王新法, 马玉琨. 基于RegNet网络的马铃薯病害识别研究[J]. 中国农机化学报, 2024, 45(6): 229-234.
[15]	钟丹, 陈鸿文, 王思, 邱霞, 蒲长兵, 李宗南. 基于MLS LiDAR点云提取桃树结构参数[J]. 中国农机化学报, 2024, 45(5): 182-187.