基于MTF-gcForest的带钢表面缺陷分类方法研究
马文杰，王杰*
（四川大学 机械工程学院，四川 成都 610065）
摘要：针对带钢表面缺陷位置分布不均、类型复杂多样的特点，为保证特征提取的维度丰富性与识别准确率，提出一种基于多纹理特征融合与gcForest集成学习相结合的带钢缺陷识别方法MTF-gcForest。首先提取带钢表面的灰度共生矩阵、局部二值模式、灰度游程矩阵特征，以充分挖掘带钢表面的纹理信息。然后，将归一化处理后的特征进行融合，最后用gcForest分类器进行分类。实验比较了单纹理特征和多纹理特征的性能表现，以及多种分类器的分类精度。实验结果表明：基于MTF-gcForest方法的平均准确率达到97.22%，优于其他带钢表面缺陷检测算法，具有较强的推广意义。
关键词：带钢；缺陷检测；纹理特征；灰度共生矩阵；灰度游程矩阵；局部二值模式；gcForest
中图分类号：TH13 ；TG316.1+92 文献标志码：A doi：10.3969/j.issn.1006-0316.2024.02.002
文章编号：1006-0316 (2024) 02-0007-07
Study On Surface Defect Classification Method for Strip Steel Based on MTF-gcForest
MA Wenjie，WANG Jie
( School of Mechanical Engineering, Sichuan University, Chengdu 610065, China )
Abstract：Aiming at the uneven distribution and complexity of strip steel surface defects, this paper proposes a strip defect classification method known as MTF-gcForest (Multi-Texture Fusion-gcForest) to ensure the dimensional richness and recognition accuracy of feature extraction. Firstly, the gray-level co-occurrence matrix (GLCM), local binary patterns (LBP), and gray-level run-length matrix (GLRLM) of the strip surface are extracted to fully excavate the texture information of the strip surface. Then, the features are normalized, fused, and finally classified with the gcForest classifier. The experiment compares the performance of single-texture feature and multi-texture feature and evaluates the classification accuracy of various classifiers. The experimental results show that the average accuracy rate based on the MTF-gcForest method reaches 97.22%, which is better than other strip surface defect detection algorithms with significant potential for widespread application.
Key words：strip steel；defect detection；texture feature；GLCM；GLRLM；LBP；gcForest
———————————————
收稿日期：2023-10-08
基金项目：四川省重点研发项目（2022YFG0058）
作者简介：马文杰（1999－），男，四川成都人，硕士研究生，主要研究方向为机械结构设计与仿真、故障检测，E-mail：mawenjie1024@stu.scu.edu.cn。*通讯作者：王杰（1964－），男，四川成都人，博士，教授，主要研究方向为计算机辅助设计与制造等，E-mail：wangjie@scu.edu.cn。
（四川大学 机械工程学院，四川 成都 610065）
摘要：针对带钢表面缺陷位置分布不均、类型复杂多样的特点，为保证特征提取的维度丰富性与识别准确率，提出一种基于多纹理特征融合与gcForest集成学习相结合的带钢缺陷识别方法MTF-gcForest。首先提取带钢表面的灰度共生矩阵、局部二值模式、灰度游程矩阵特征，以充分挖掘带钢表面的纹理信息。然后，将归一化处理后的特征进行融合，最后用gcForest分类器进行分类。实验比较了单纹理特征和多纹理特征的性能表现，以及多种分类器的分类精度。实验结果表明：基于MTF-gcForest方法的平均准确率达到97.22%，优于其他带钢表面缺陷检测算法，具有较强的推广意义。
关键词：带钢；缺陷检测；纹理特征；灰度共生矩阵；灰度游程矩阵；局部二值模式；gcForest
中图分类号：TH13 ；TG316.1+92 文献标志码：A doi：10.3969/j.issn.1006-0316.2024.02.002
文章编号：1006-0316 (2024) 02-0007-07
Study On Surface Defect Classification Method for Strip Steel Based on MTF-gcForest
MA Wenjie，WANG Jie
( School of Mechanical Engineering, Sichuan University, Chengdu 610065, China )
Abstract：Aiming at the uneven distribution and complexity of strip steel surface defects, this paper proposes a strip defect classification method known as MTF-gcForest (Multi-Texture Fusion-gcForest) to ensure the dimensional richness and recognition accuracy of feature extraction. Firstly, the gray-level co-occurrence matrix (GLCM), local binary patterns (LBP), and gray-level run-length matrix (GLRLM) of the strip surface are extracted to fully excavate the texture information of the strip surface. Then, the features are normalized, fused, and finally classified with the gcForest classifier. The experiment compares the performance of single-texture feature and multi-texture feature and evaluates the classification accuracy of various classifiers. The experimental results show that the average accuracy rate based on the MTF-gcForest method reaches 97.22%, which is better than other strip surface defect detection algorithms with significant potential for widespread application.
Key words：strip steel；defect detection；texture feature；GLCM；GLRLM；LBP；gcForest
———————————————
收稿日期：2023-10-08
基金项目：四川省重点研发项目（2022YFG0058）
作者简介：马文杰（1999－），男，四川成都人，硕士研究生，主要研究方向为机械结构设计与仿真、故障检测，E-mail：mawenjie1024@stu.scu.edu.cn。*通讯作者：王杰（1964－），男，四川成都人，博士，教授，主要研究方向为计算机辅助设计与制造等，E-mail：wangjie@scu.edu.cn。
- 上一篇：冲击载荷下某船用起重机贮存强度评估
- 下一篇：基于改进鲸鱼算法的搬运机器人运动规划
设为首页 | 加入收藏 | 免责条款
《机械》杂志版权所有 Copyright©2008-2012 Jixiezazhi.com All Rights Reserved
电话：028-85925070 传真：028-85925073 E-mail：jixie@vip.163.com
地址：四川省成都锦江工业开发区墨香路48号 邮编：610063
Powered by PageAdmin CMS