Grain size grading method based on improved UNet network and optimized algorithm of grain boundary

doi:10.13251/j.issn.0254-6051.2024.07.025

Abstract

Abstract: Grain size has an undeniable impact on the properties of metal materials, and the manual grading methods of grain size is difficult to meet the current detection needs of metal materials. Therefore, for austenite microstructure, a grain size automatic grading method based on improved UNet network and grain boundary optimized algorithm was proposed, and the accuracy of the austenite grading results calculated by the method was analyzed by comparing with the manual grading results. The results show that when the improved UNet network is used to segment austenite grain boundaries, and then the Hough transform based grain boundary optimization algorithm is used to detect and remove twin grain boundaries and branch burr, the grain boundary extraction effect can be effectively optimized, and the accuracy of subsequent grain size calculation can be improved. The absolute error between the austenite grading results obtained by the proposed algorithm and that by the manual method is within 0.25, indicating that the algorithm can efficiently, conveniently and accurately complete the grading of austenite grain size.

Key words: UNet network, Hough transform, optimized algorithm, grain size grading

CLC Number:

TF31

Qi Xueqian, Huang Xiaohong, Song Yue, Liu Yanping, Zhang Luyue, Zhang Qingjun. Grain size grading method based on improved UNet network and optimized algorithm of grain boundary[J]. Heat Treatment of Metals, 2024, 49(7): 161-167.

References

[1]刘文仲. 中国钢铁工业智能制造现状及思考[J]. 中国冶金, 2020, 30(6): 1-7.
Liu Wenzhong. Current situation and thinking of intelligent manufacturing in China's iron and steel industry[J]. China Metallurgy, 2020, 30(6): 1-7.
[2]Zhang Lixin, Xu Zhengguang, Wei Shuailing, et al. Grain size automatic determination for 7050Al alloy based on a fuzzy logic method[J]. Rare Metal Materials and Engineering, 2016, 45(3): 548-554.
[3]刁福林, 陈昊, 李延丰, 等. 高温合金的晶界提取及特征定量表征方法[J]. 失效分析与预防, 2019, 14(1): 25-32.
Diao Fulin, Chen Hao, Li Yanfeng, et al. Study on grain boundary extraction of alloy materials and its quantitative evaluation method[J]. Failure Analysis and Prevention, 2019, 14(1): 25-32.
[4]朱建栋. 钢材金相图像晶界提取算法研究及智能评级软件设计[D]. 镇江: 江苏大学, 2018.
Zhu Jiandong. Research on the algorithm used for grain boundary extraction from metallographic images and design of intelligent rating software[D]. Zhenjiang: Jiangsu University, 2018.
[5]许帧英, 张琦, 朱建栋. 基于纹理特征的集成学习金相自动评级方法[J]. 信息技术, 2018(9): 71-74, 78.
Xu Zhenying, Zhang Qi, Zhu Jiandong. Metallographic automatic rating method based on ensemble learning with texture features[J]. Information Technology, 2018(9): 71-74, 78.
[6]孟祥亮. 图像识别技术在S355J2W钢板带状组织评定中的应用[J]. 理化检验(物理分册), 2019, 55(9): 598-601, 606.
Meng Xiangliang. Application of image recognition technology in evaluation of banded structure of S4355J2W steel plate[J]. Physical Testing and Chemical Analysis (Part A: Physical Testing), 2019, 55(9): 598-601, 606.
[7]韩越祥, 卢振发. 基于图像处理20钢的晶界提取算法[J]. 实验室研究与探索, 2021, 40(8): 32-36.
Han Yuexiang, Lu Zhenfa. Research on the algorithm of grain boundary extraction of 20 steel based on image processing[J]. Research and Exploration in Laboratory, 2021, 40(8): 32-36.
[8]张佳宁. 基于深度学习的金相晶粒度自动评级系统[D]. 北京: 冶金自动化研究设计院, 2021.
Zhang Jianing. Automatic grain size rating of metallographic based on deep learning[D]. Beijing: Automation Research and Design Institute of Metallurgical Industry, 2021.
[9]张利欣, 孙涵, 尧昊天, 等. 基于自适应标记的金相组织智能检测方法[J]. 实验室研究与探索, 2022, 41(8): 1-4, 101.
Zhang Lixin, Sun Han, Yao Haotian, et al. Research on metallographic structure detection method based on adaptive marked[J]. Research and Exploration in Laboratory, 2022, 41(8): 1-4, 101.
[10]刘玠. 人工智能推动冶金工业变革[J]. 钢铁, 2020, 55(6): 1-7.
Liu Jie. Artificial intelligence drives changes in metallurgical industry[J]. Iron and Steel, 2020, 55(6): 1-7.
[11]立济伟. 基于深度学习的锂电池极片缺陷检测与分类系统[D]. 上海: 上海电机学院, 2022.
Li Jiwei. Lithium battery pole piece defect detection and classification system based on deep learning[D]. Shanghai: Shanghai Dianji University, 2022.
[12]黄晓鸣, 何富运, 唐晓虎, 等. U-Net及其变体在医学图像分割中的应用研究综述[J]. 中国生物医学工程学报, 2022, 41(5): 567-576.
Huang Xiaoming, He Fuyun, Tang Xiaohu, et al. Review on applications of U-Net and its variants in medical image segmentation[J]. Chinese Journal of Biomedical Engineering, 2022, 41(5): 567-576.
[13]徐思则, 刘威. 基于UNet网络的乳腺癌肿瘤细胞图像分割[J]. 电子设计工程, 2022, 30(12): 63-66, 73.
Xu Size, Liu Wei. UNet-based image segmentation of breast cancer tumor cells[J]. Electronic Design Engineering, 2022, 30(12): 63-66, 73.
[14]周思雨, 储岳中, 张学锋. 基于改进UNet网络的烧结矿气孔分割[J]. 计算机系统应用, 2022, 31(7): 278-284.
Zhou Siyu, Chu Yuezhong, Zhang Xuefeng. Segmentation of sinter pores based on improved UNet network[J]. Computer Systems and Applications, 2022, 31(7): 278-284.
[15]李原, 李燕君, 刘进超, 等. 基于改进Res-UNet网络的钢铁表面缺陷图像分割研究[J]. 电子与信息学报, 2022, 44(5): 1513-1520.
Li Yuan, Li Yanjun, Liu Jinchao, et al. Research on segmentation of steel surface defect images based on improved Res-UNet network[J]. Journal of Electronics and Information Technology, 2022, 44(5): 1513-1520.
[16]宋翰凌, 孟晓亮, 罗森, 等. 基于深度学习的连铸坯低倍质量评级[J]. 冶金自动化, 2023, 47(2): 73-81.
Song Hanlin, Meng Xiaoliang, Luo Sen, et al. Evaluation on macro quality of strand based on deep learning method[J]. Metallurgical Industry Automation, 2023, 47(2): 73-81.
[17]许涛, 麻爱松, 吕欢, 等. 基于改进UNet模型的原棉杂质图像分割方法[J]. 西安工程大学学报, 2023, 37(1): 77-83.
Xu Tao, Ma Aisong, Lü Huan, et al. A raw cotton impurity image segmentation method based on improved UNet model[J]. Journal of Xi'an Polytechnic University, 2023, 37(1): 77-83.
[18]郭金涛, 王龙, 余建波, 等. 基于深度学习的宽厚板热轧轧制力预测[J]. 锻压技术, 2022, 47(7): 167-174.
Guo Jintao, Wang Long, Yu Jianbo, et al. Prediction on rolling force in hot rolling of wide and thick plate based on deep learning[J]. Forging and Stamping Technology, 2022, 47(7): 167-174.
[19]程金浪. 快速并行细化算法检测地震反射同相轴[J]. 内蒙古煤炭经济, 2019(5): 156-157, 145.
Cheng Jinlang. Research on fast parallel thinning algorithm for detecting seismic reflection coaxial[J]. Inner Mongolia Coal Economy, 2019(5): 156-157, 145.
[20]张梦原, 陈俊智. 基于数字图像处理的岩体结构面迹线检测[J]. 化工矿物与加工, 2023, 52(1): 29-33.
Zhang Mengyuan, Chen Junzhi. Rock mass discontinuity trace detection based on digital image processing[J]. Industrial Minerals and Processing, 2023, 52(1): 29-33.
[21]GB/T 6394—2017, 金属平均晶粒度测定方法[S].