摘要
传统的机械零部件表面硬度检测方法通常需要与零部件表面直接接触,会对零部件表面造成损伤,影响检测效果。因此,提出基于机器视觉的机械零部件表面硬度检测研究。机器视觉技术能捕捉机械零部件表面的大量图像信息,助力硬度检测。利用机器视觉技术对图像进行预处理,通过灰度值分析和动态阈值确定,提取机械零部件表面异常硬度特征。构建硬度评估模型,综合考虑异常硬度特征对硬度值的贡献和多个异常硬度特征对硬度值的综合影响,通过训练和校准确定权重系数,实现机械零部件表面硬度的快速准确检测。结果表明,设计方法通过纹理特征量化材料表面硬度,平均相关系数达0.876,在部件合格率方面,设计方法在样本11时达95%,验证了其高准确性和可靠性。
The traditional surface hardness testing method for mechanical components usually requires direct contact with the component surface,which can cause damage to the component surface and affect the testing effect.Therefore,a research on surface hardness detection of mechanical components based on machine vision is proposed.Machine vision technology can capture a large amount of image information on the surface of mechanical components,assisting in hardness testing.Machine vision technology is used to preprocess images,extracting abnormal hardness features on the surface of mechanical components through grayscale value analysis and dynamic threshold determination.A hardness evaluation model is built that comprehensively considers the contribution of abnormal hardness features to hardness values and the comprehensive influence of multiple abnormal hardness features on hardness values.Determine weight coefficients through training and calibration to achieve rapid and accurate detection of surface hardness of mechanical components.The results showed that the design method quantified the surface hardness of materials through texture features,with an average correlation coefficient of 0.876.In terms of component qualification rate,the design method achieved 95%in sample 11,verifying its high accuracy and reliability.
作者
李宗禛
LI Zongzhen(School of Mechanical and Electrical Engineering,Shandong Jianzhu University,Jinan,Shandong 250101,China)
出处
《自动化应用》
2025年第14期59-62,共4页
Automation Application
关键词
机器视觉
CMOS相机
机械零部件
表面硬度
异常硬度特征
machine vision
CMOS camera
mechanical components
surface hardness
abnormal hardness characteristics
