基于灰度共生矩阵的图形纹理检测及焊接缺陷的SVM分类实现  

SVM Classification Implementation of Graphic Texture Detection and Welding Defects Based on Gray Level Co-occurrence Matrix

作　　者：陈滔 张庆国 刘澳 CHEN Tao;ZHANG Qingguo;LIU Ao(School of Engineering,Anhui Agricultural University,Hefei 230036,China;School of Civil and Commercial Economic Law,Gansu University of Political Science and Law,Lanzhou 730070,China;School of Clinical Medicine,Anhui Medical University,Hefei 230031,China;School of Information&Computer,Anhui Agricultural University,Hefei 230036,China)

机构地区：[1]安徽农业大学工学院,安徽合肥230036 [2]甘肃政法大学民商经济法学院,甘肃兰州720070 [3]安徽医科大学临床医学院,安徽合肥230031 [4]安徽农业大学信息与计算机学院,安徽合肥230036

出　　处：《洛阳理工学院学报（自然科学版）》2022年第1期53-61,67,共10页Journal of Luoyang Institute of Science and Technology：Natural Science Edition

基　　金：安徽农业大学“优才计划”科研发展资助项目(XSZZ202006);安徽省学术和技术带头人及后备人选学术科研活动经费(2016H072).

摘　　要：工业4.0时代,焊接技术作为一种最基本的工件加工技术,被广泛应用于各种工业生产。焊接质量直接影响焊接产品的使用寿命,从而影响工业生产活动的效率。基于灰度共生矩阵(GLCM)对X-射线焊接缺陷图像进行特征提取,分析X-射线焊接缺陷的分类特点,构建SVM多类分类器,分析对比不同核函数对分类精度的影响。基于RBF核函数的SVM分类器能够对焊接缺陷进行良好的识别分类,总体分类精度达到了92.6%,为焊接缺陷的检测识别提供了一种简便的方法。In the era of industry 4.0,as the most basic workpiece processing technology,welding is widely used in various industrial production environments.The quality of welding directly affects the service life of welding products,thus affecting the efficiency of industrial production activities.With the feature extraction of X-ray welding defect image based on,the paper analyzes the classification characteristics of X-ray welding defects,constructs a SVM multi class classifier,and compares the influence of different kernel functions on the classification accuracy.The results are that the SVM classifier based on RBF kernel function can better realize the defect identification and classification,and the overall classification accuracy reaches 92.6%.It provides a simple method for the detection and identification of welding defects.

关 键 词：缺陷识别 特征提取 SVM分类 核函数 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] TG441.7[自动化与计算机技术—计算机科学与技术]