摘要
针对接触网旋转双耳销钉,提出一种基于环形对称Gabor变换特征的销钉松脱与缺失状态检测方法。融合旋转不变LBP特征与HOG特征的方法训练SVM分类器,实现旋转双耳定位识别;采用圆弧检测完成销钉准确定位,实现销钉区域的分割;利用环形对称Gabor变换完成纹理边缘信息的特征提取,结合BP神经网络实现对旋转双耳销钉状态的判断识别。研究和实验结果表明,该方法实现了在销钉穿插方向任意和形状多样等复杂情况下的状态识别。
For the detection of catenary rotating binaural pins,this paper proposes a method used to detect the conditions of the pins based on CSGT.In this method,the rotation invariant LBP is used with HOG to train the linear SVM classifier and identify the pins.The circular inspection method is used to accurately position the pins and split its area and CSGT is used to accomplish the feature extraction of the edge Information.Then,the BP neural network is used to judge and identify the condition of the rotating binaural pins.The research and experiment results show that the method can be used to accurately identify the position of the catenary rotating binaural pins.
作者
高聪
林建辉
邓韬
杨见光
GAO Cong;LIN Jianhui;DENG Tao;YANG Jianguang(State-Key Laboratory of Traction Power,Southwest Jiaotong University,Chengdu 610031,China)
出处
《机械制造与自动化》
2020年第2期166-169,共4页
Machine Building & Automation
基金
国家自然科学基金项目(51475387)。
