图像识别智能化处理技术在塑料齿轮缺齿检测中的应用
|
摘要
基于图像识别智能化处理技术,研究了其在塑料齿轮缺齿检测中的应用。结果表明:通过图像融合技术可融合同类齿轮缺齿源图像,综合这些特征可提高齿轮缺齿识别率;通过WTDS实验平台对齿轮缺齿振动信号进行采集,经过小波包双谱分析后,生成的双谱图纹理均有各自的特征,通过融合可对不同通道信息进行综合,有利于诊断齿轮缺齿故障;通过相关性分析可将双谱图GLCM矩阵的12维特征向量降为5维,数据冗余度得到大幅降低,机器学习效率得到提高。最后采用支持向量机对基于图像特征的齿轮缺齿进行诊断,其中径向基核函数识别率最高,为93.75%。
Based on image recognition and intelligent processing technology, the application of this technology in the detection of plastic gear teeth missing was studied. The results show that the image fusion technology can fuse the image of the same kind of gear teeth missing source, and the recognition rate of gear teeth missing can be improved by synthesizing these features. The vibration signal of gear tooth missing is collected by WTDS experimental platform. After the analysis of wavelet packet bispectrum, the texture of the bispectrum has its own characteristics. The information of different channels can be synthesized by fusion, which is helpful for the diagnosis of gear tooth missing fault. By correlation analysis, the 12-dimensional eigenvector of GLCM matrix of bispectrum can be reduced to 5-dimensional, the redundancy of data can be greatly reduced, and the efficiency of machine learning can be improved. Finally, support vector machine is used to diagnose gear tooth missing based on image features. The recognition rate of radial basis function(RBF) is the highest, which is 93.75%.
Based on image recognition and intelligent processing technology, the application of this technology in the detection of plastic gear teeth missing was studied. The results show that the image fusion technology can fuse the image of the same kind of gear teeth missing source, and the recognition rate of gear teeth missing can be improved by synthesizing these features. The vibration signal of gear tooth missing is collected by WTDS experimental platform. After the analysis of wavelet packet bispectrum, the texture of the bispectrum has its own characteristics. The information of different channels can be synthesized by fusion, which is helpful for the diagnosis of gear tooth missing fault. By correlation analysis, the 12-dimensional eigenvector of GLCM matrix of bispectrum can be reduced to 5-dimensional, the redundancy of data can be greatly reduced, and the efficiency of machine learning can be improved. Finally, support vector machine is used to diagnose gear tooth missing based on image features. The recognition rate of radial basis function(RBF) is the highest, which is 93.75%.
引文
[1]高忠生.计算机图像识别的智能化处理技术分析[J].电脑知识与技术,2013,9(15):3 591-3 593.
[2]吉文帅.计算机图像识别智能化处理技术的分析[J].山东工业技术,2017(15):100.
[3]任楷飞,孟令军,顾泽凌.基于Lab VIEW的塑料齿轮缺齿检测系统设计[J].现代电子技术,2018,41(4):88-92.
[4]谢凤英,赵丹培,李露,等.数字图像处理及应用[M].北京:电子工业出版社,2014.
[5]Lin C,Cao X J,Fan Y,et al.Pitch deviation measurement and analysis of curve-face gearpair[J].Measurement,2016(81):95-101.
[6]苏卫.齿轮传动装置的故障诊断仪设计[D].兰州:兰州理工大学,2014.
[7]施保华,魏雅慧.一种基于机器视觉的螺纹缺陷检测方法[J].三峡大学学报:自然科学版,2015,37(2):64-66.
[8]张澎涛.齿轮箱故障振动信号去噪及特征提取算法研究[D].哈尔滨:东北林业大学,2014.
[9]Wei Q,Dingguo L.A survey on wind turbine condition monitoring and fault diagnosis.Part I:Components and subsystems[J].IEEE Transactions on Industrial Electronics,2015,62(10):6 536-6 545.
[10]章浙涛,朱建军,匡翠林,等.小波包多阈值去噪法及其在形变分析中的应用[J].测绘学报,2014(1):13-20.
[2]吉文帅.计算机图像识别智能化处理技术的分析[J].山东工业技术,2017(15):100.
[3]任楷飞,孟令军,顾泽凌.基于Lab VIEW的塑料齿轮缺齿检测系统设计[J].现代电子技术,2018,41(4):88-92.
[4]谢凤英,赵丹培,李露,等.数字图像处理及应用[M].北京:电子工业出版社,2014.
[5]Lin C,Cao X J,Fan Y,et al.Pitch deviation measurement and analysis of curve-face gearpair[J].Measurement,2016(81):95-101.
[6]苏卫.齿轮传动装置的故障诊断仪设计[D].兰州:兰州理工大学,2014.
[7]施保华,魏雅慧.一种基于机器视觉的螺纹缺陷检测方法[J].三峡大学学报:自然科学版,2015,37(2):64-66.
[8]张澎涛.齿轮箱故障振动信号去噪及特征提取算法研究[D].哈尔滨:东北林业大学,2014.
[9]Wei Q,Dingguo L.A survey on wind turbine condition monitoring and fault diagnosis.Part I:Components and subsystems[J].IEEE Transactions on Industrial Electronics,2015,62(10):6 536-6 545.
[10]章浙涛,朱建军,匡翠林,等.小波包多阈值去噪法及其在形变分析中的应用[J].测绘学报,2014(1):13-20.