摘要
针对传统受电弓碳滑板检测中检测效率低、检测精度差等缺点,提出一种基于Mask R-CNN的优化改进算法。该算法采用铁道部受电弓损坏评定的新规定及实地的样本数据集,通过改进特征提取算法的网络结构以及优化损失值来提高算法对图像的处理效率,实现受电弓碳滑板缺陷的掩膜准确标注,有效减小受电弓滑板的损毁对电力机车运行的影响。最终通过实验验证该算法对受电弓碳滑板缺陷的检测精度和效率有明显的提升作用。
In this paper,an optimized and improved algorithm based on Mask R-CNN is proposed to solve the shortcomings of traditional pantograph carbon slide detection,such as low detection efficiency and poor detection accuracy.The algorithm adopted the ministry of railways pantograph damage assessment of the new regulations and field of the sample data sets.By improving feature extraction algorithm of network structure and optimizing the loss value,the efficiency of the algorithm of image processing was improved,which realized the pantograph slide carbon defects mask label accurately and reduced the loss of the pantograph slide effects on electric locomotive running.Experimental results show that this algorithm can improve the detection accuracy and efficiency of pantograph carbon slide.
作者
韩璐
刘太豪
宋海亮
宋佳
Han Lu;Liu Taihao;Song Hailiang;Song Jia(School of Electrical Information,Southwest Petroleum University,Chengdu 610500,Sichuan,China)
出处
《计算机应用与软件》
北大核心
2024年第1期105-111,176,共8页
Computer Applications and Software
基金
国家自然科学基金项目(51607151)。
