MSSIF-Net: an efficient CNN automatic detection method for freight train images
MSSIF-Net: an efficient CNN automatic detection method for freight train images
复制标题
DOI：
10.1007/s00521-022-08035-1
复制
发表时间：
2022-11
影响因子：
6
通讯作者：
Longxin Zhang;Yang Hu;Jingsheng Chen;Chuang Li;Keqin Li
中科院分区：
文献类型：
--
作者：
Longxin Zhang;Yang Hu;Jingsheng Chen;Chuang Li;Keqin Li
货运列车是最重要的运输方式之一。货物列车部件的故障检测是保证列车运行安全的关键。针对传统列车故障检测方法检测效率低、准确率低的问题，提出了一种基于YOLOv 4的多尺度空间信息融合CNN网络（MSSIF-Net）的单阶段目标检测方法。采用自适应空间特征融合方法和多尺度通道注意机制，构建多尺度特征共享网络，实现不同层次的特征信息共享，提高检测精度。MSSIF-Net在训练图像测试集、PASCAL VOC 2007测试集和表面缺陷检测数据集上的平均精度值分别为94.73%、87.76%和75.54%，优于YOLOv 4、Faster R-CNN、CenterNet、RetinaNet和YOLOX-1。MSSIF-Net的检测速度为33.10 FPS，在检测精度和速度之间实现了很好的平衡。此外，MSSIF-Net的性能估计后，添加噪声或旋转的火车图像在一个轻微的角度，以模拟一个真实的场景。实验结果表明，MSSIF-Net具有良好的抗干扰能力。
Freight trains are one of the most important modes of transportation. The fault detection of freight train parts is crucial to ensure the safety of train operation. Given the low detection efficiency and accuracy of traditional train fault detection methods, a novel one-stage object detection method called the multi-scale spatial information fusion CNN network (MSSIF-Net) based on YOLOv4 is proposed in this study. The adaptive spatial feature fusion method and multi-scale channel attention mechanism are used to construct the multi-scale feature sharing network and consequently realize feature information sharing at different levels and promote detection accuracy. The mean average precision values of MSSIF-Net on the train image test set, PASCAL VOC 2007 test set, and surface defect detection dataset are 94.73%, 87.76%, and 75.54%, respectively, outperforming YOLOv4, Faster R-CNN, CenterNet, RetinaNet, and YOLOX-l. The detection speed of MSSIF-Net is 33.10 FPS, achieving a good balance between detection accuracy and speed. In addition, the MSSIF-Net performance is estimated after adding noise or rotating the train images at a slight angle to simulate a real scene. Experimental results indicate that MSSIF-Net has favorable anti-interference ability.