摘要
针对金属产品上的缺陷在识别过程中,缺陷种类繁多、不易分类等问题,采用一种改进灰狼优化算法优化SVM核参数的思想,并构建改进的GWO-SVM分类器模型实现金属产品表面缺陷识别。首先,在缺陷区域分割的基础上,通过LBP算法对金属产品零件的缺陷数据进行特征提取;其次,通过主成分分析算法对数据进行降维处理;最后,采用SVM对数据样本进行分类识别。实验表明,与其他的分类器模型相比较,文中所设计的改进分类器模型更能够精确有效地对零件不同形状缺陷进行分类识别。
The work aims to aiming at the problems ofthe wide variety and difficult classification of defects in metal products during the identification process,an improved GWO algorithm is adopted to optimize the kernel parameters of SVM,and an improved GWO-SVM classifier model is constructed to realize the surface defect identification of metal products.Firstly,based on the defect region segmentation,the feature of the defect data of metal products is extracted by LBP algorithm,and the dimension of the data is reduced by principal component analysis algorithm.Finally,SVM is used to classify and identify the data samples.In order to overcome the local extremum and poor convergence precision of SVM penalty factors and kernel function parameters,an improved GWO algorithm is designed to optimize the relevant parameters of SVM to improve the defect recognition rate of metal products.Experimental results showed that compared with other classifier models,the improved classifier model designed in this paper can classify and recognize different shape defects more accurately and effectively,which provides a favourable guarantee for improving the surface defect recognition of metal products.
作者
杨益服
李文磊
李俊杰
问轲
张炜
YANG Yi-fu;LI Wen-lei;LI Jun-jie;WEN Ke;ZHANG Wei(Zhejiang Hoping Machinery Co.,Ltd.,Wenzhou 325003,China;Qing Gong College,Harbin University of Commerce,Harbin 150028,China)
出处
《组合机床与自动化加工技术》
北大核心
2022年第11期143-146,150,共5页
Modular Machine Tool & Automatic Manufacturing Technique