[1]黄贤山,卢 冶,张小立,等.基于迁移学习的灰铸铁金相组织分类研究[J].计算机技术与发展,2021,31(增刊):21-25.[doi:10. 3969 / j. issn. 1673-629X. 2021. S. 004]
　HUANG Xian-shan,LU Ye,ZHANG Xiao-li,et al.Research on Classification of Gray Cast Iron Metallographic Structure Based on Transfer Learning[J].,2021,31(增刊):21-25.[doi:10. 3969 / j. issn. 1673-629X. 2021. S. 004]
点击复制

基于迁移学习的灰铸铁金相组织分类研究()

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
31
期数:
2021年增刊
页码:
21-25
栏目:
人工智能
出版日期:
2021-12-31

文章信息/Info

Title:
Research on Classification of Gray Cast Iron Metallographic Structure Based on Transfer Learning
文章编号:
1673-629X(2021)S0021-05
作者:
黄贤山1 ; 卢 冶1 ; 张小立2 ; 杨 俊2
1. 江苏科技大学 电气与信息工程学院,江苏 张家港 215600;
2. 江苏科技大学 冶金与材料工程学院,江苏 张家港 215600
Author(s):
HUANG Xian-shan1 ; LU Ye1 ; ZHANG Xiao-li2 ; YANG Jun2
1. School of Electrical and Information Engineering,Jiangsu University of Science and Technology,Zhangjiagang 215600,China;
2. School of Metallurgy and Materials Engineering,Jiangsu University of Science and Technology,Zhangjiagang 215600,China
关键词:
深度学习; 迁移学习; 灰铸铁; 卷积神经网络; VGG16 模型; ResNet50 模型
Keywords:
deep learning; transfer learning; gray cast iron; convolutional neural network; VGG16 model; ResNet50 model
分类号:
TP18;TG115
DOI:
10. 3969 / j. issn. 1673-629X. 2021. S. 004
摘要:
针对灰铸铁金相组织分类问题,提出了基于卷积神经网络和迁移学习的判别方法。 首先收集在显微镜下放大 100倍后的金相组织图像数据,再利用直方图均衡化和高斯滤波等算法对原始图像进行预处理,将处理后的图像数据分为训练集和测试集;然后对已经在源领域训练过的卷积神经网络模型进行微调,并使用经过数据增强的训练集数据对微调后的模型进行再次训练,得到模型最优的参数;最后利用测试集数据评估模型的泛化能力,在实验使用的两种卷积神经网络模型 VGG16 和 ResNet50 上,测试集的预测准确率分别达到了 87. 6% 和 93. 8% 。 该结果表明在灰铸铁金相组织分类的问题上,基于卷积神经网络,并结合迁移学习的方法具有较好的特征提取能力和分类能力,该方法效率高、准确率高,且不需要人为干预,在工业生产中具有广泛的应用前景。
Abstract:
To classify images of gray cast iron metallographic structure,a discrimination method based on convolutional neural network and transfer learning is proposed. Collecting the metallographic tissue image data magnified 100 times under the microscope, and the original images are preprocessed by algorithms such as histogram equalization and Gaussian filtering,after that,a training set and a test set are split apart. Then we fine-tune the pre-trained convolutional neural network model,which has been trained in the source field and dodata augmentation on training set,by which the fine-tuned model is retrained to get the optimal parameters. Finally,the test set is used to evaluate the generalization ability of the model. In the experiment, we apply two widely used convolutional neural network modelsVGG16 and ResNet50, the prediction accuracy of the test set reaches 87. 6% and 93. 8% . It is shown that on the problem of metallographic structure classification of cast iron,the method based on the convolutional neural network and combined with the transfer learning has better feature extraction and classification capabilities. This method has high efficiency and high accuracy,mean while it does not require human intervention,which has broad application prospects in industrial production.

相似文献/References:

[1]陈浩翔,蔡建明,刘铿然,等. 手写数字深度特征学习与识别[J].计算机技术与发展,2016,26(07):19.
　CHEN Hao-xiang,CAI Jian-ming,LIU Keng-ran,et al. Deep Learning and Recognition of Handwritten Numeral Features[J].,2016,26(增刊):19.
[2]贺飞翔,赵启军. 基于深度学习的头部姿态估计[J].计算机技术与发展,2016,26(11):1.
　HE Fei-xiang,ZHAO Qi-jun. Head Pose Estimation Based on Deep Learning[J].,2016,26(增刊):1.
[3]曾志平[] [],萧海东[],张新鹏[]. 基于DBN的金融时序数据建模与决策[J].计算机技术与发展,2017,27(04):1.
　ZENG Zhi-ping[] [],XIAO Hai-dong[],ZHANG Xin-peng[]. Modeling and Decision-making of Financial Time Series Data with DBN[J].,2017,27(增刊):1.
[4]李宏林. 分析式纹理合成技术及其在深度学习的应用[J].计算机技术与发展,2017,27(11):7.
　LI Hong-lin. Analyzed Texture-synthesis Techniques and Their Applications in Deep Learning[J].,2017,27(增刊):7.
[5]曲之琳,胡晓飞.基于改进激活函数的卷积神经网络研究[J].计算机技术与发展,2017,27(12):77.[doi:10.3969/ j. issn.1673-629X.2017.12.017]
　QU Zhi-lin,HU Xiao-fei.Research on Convolutional Neural Network Based on Improved Activation Function[J].,2017,27(增刊):77.[doi:10.3969/ j. issn.1673-629X.2017.12.017]
[6]吴 超,邵 曦.基于深度学习的指静脉识别研究[J].计算机技术与发展,2018,28(02):200.[doi:10．3969/j．issn．1673－629X．2018．02．043]
　WU Chao,SHAO Xi.Research on Finger Vein Ｒecognition Based on Deep Learning[J].,2018,28(增刊):200.[doi:10．3969/j．issn．1673－629X．2018．02．043]
[7]陈强锐,谢世朋.基于深度学习的肺部肿瘤检测方法[J].计算机技术与发展,2018,28(04):201.[doi:10.3969/ j. issn.1673-629X.2018.04.043]
　CHEN Qiang-rui,XIE Shi-peng.Lung Cancer Detection Method Based on Deep Learning[J].,2018,28(增刊):201.[doi:10.3969/ j. issn.1673-629X.2018.04.043]
[8]施泽浩,赵启军.基于全卷积网络的目标检测算法[J].计算机技术与发展,2018,28(05):55.[doi:10.3969/j.issn.1673－629X.2018.05.013]
　SHI Ze-hao,ZHAO Qi-jun.Object Detection Algorithm Based on Fully Convolutional Neural Network[J].,2018,28(增刊):55.[doi:10.3969/j.issn.1673－629X.2018.05.013]
[9]高翔,陈志,岳文静,等.基于视频场景深度学习的人物语义识别模型[J].计算机技术与发展,2018,28(06):53.[doi:10.3969/ j. issn.1673-629X.2018.06.012]
　GAO Xiang,CHEN Zhi,YUE Wen-jing,et al.Human Semantic Recognition Model Based on Video Scene Deep Learning[J].,2018,28(增刊):53.[doi:10.3969/ j. issn.1673-629X.2018.06.012]
[10]王嘉文,王传栋,杨雁莹.一种中文人名识别的训练架构[J].计算机技术与发展,2018,28(07):53.[doi:10.3969/ j. issn.1673-629X.2018.07.012]
　WANG Jia-wen,WANG Chuan-dong,YANG Yan-ying.A Training Framework for Chinese Name Recognition[J].,2018,28(增刊):53.[doi:10.3969/ j. issn.1673-629X.2018.07.012]
[11]蒋 鹏,何 勇,姚凯学,等.基于深度学习的糖尿病眼底病变分级方法研究[J].计算机技术与发展,2021,31(12):193.[doi:10. 3969 / j. issn. 1673-629X. 2021. 12. 032]
　JIANG Peng,HE Yong,YAO Kai-xue,et al.Research on Classification of Diabetic Retinopathy Based on Deep Learning[J].,2021,31(增刊):193.[doi:10. 3969 / j. issn. 1673-629X. 2021. 12. 032]
[12]陈 鑫,叶 宁,徐 康,等.基于 EfficientNet 模型的毫米波雷达人体行为识别[J].计算机技术与发展,2022,32(09):134.[doi:10. 3969 / j. issn. 1673-629X. 2022. 09. 021]
　CHEN Xin,YE Ning,XU Kang,et al.FMCW Radar Human Action Recognition System Based on EfficientNet Model[J].,2022,32(增刊):134.[doi:10. 3969 / j. issn. 1673-629X. 2022. 09. 021]
[13]林泽阳,赖 俊,陈希亮.基于课程学习的深度强化学习研究综述[J].计算机技术与发展,2022,32(11):16.[doi:10. 3969 / j. issn. 1673-629X. 2022. 11. 003]
　LIN Ze-yang,LAI Jun,CHEN Xi-liang.An Overview of Deep Reinforcement Learning Based on Curriculum Learning[J].,2022,32(增刊):16.[doi:10. 3969 / j. issn. 1673-629X. 2022. 11. 003]
[14]张 睿,李允臣,王家宝*,等.基于深度学习的红外目标检测综述[J].计算机技术与发展,2023,33(11):1.[doi:10. 3969 / j. issn. 1673-629X. 2023. 11. 001]
　ZHANG Rui,LI Yun-chen,WANG Jia-bao*,et al.Survey on Infrared Object Detection Based on Deep Learning[J].,2023,33(增刊):1.[doi:10. 3969 / j. issn. 1673-629X. 2023. 11. 001]
[15]李 林,俞 璐,韩昌芝,等.多源域适应方法综述[J].计算机技术与发展,2024,34(03):1.[doi:10. 3969 / j. issn. 1673-629X. 2024. 03. 001]
　LI Lin,YU Lu,HAN Chang-zhi,et al.A Review of Multi-source Domain Adaptation[J].,2024,34(增刊):1.[doi:10. 3969 / j. issn. 1673-629X. 2024. 03. 001]
[16]李子俊,李涛*,陈浩东,等.基于抽象语法树特征迁移的软件漏洞检测方法(AST-FMVD)[J].计算机技术与发展,2024,34(06):81.[doi:10.20165/j.cnki.ISSN1673-629X.2024.0065]
　LI Zi-jun,LI Tao*,CHEN Hao-dong,et al.Software Vulnerability Detection Method Based on Abstract Syntax Tree Feature Migration(AST-FMVD)[J].,2024,34(增刊):81.[doi:10.20165/j.cnki.ISSN1673-629X.2024.0065]

更新日期/Last Update: 2021-09-10