Foreign object detection for railway ballastless trackbeds: A semisupervised learning method

被引：17

作者：

Chen, Zhengxing ^{[1
,2
]}

Wang, Qihang ^{[1
,2
]}

Yu, Tianle ^{[3
]}

Zhang, Min ^{[4
]}

Liu, Qibin ^{[4
]}

Yao, Jidong ^{[3
]}

Wu, Yanhua ^{[3
]}

Wang, Ping ^{[1
,2
]}

He, Qing ^{[1
,2
]}

机构：

[1] Southwest Jiaotong Univ, Key Lab High Speed Railway Engn, Minist Educ, Chengdu 610031, Peoples R China

[2] Southwest Jiaotong Univ, Sch Civil Engn, Chengdu 610031, Peoples R China

[3] Shanghai Dongfang Maritime Engn Technol Co Ltd, 88 Xiangyun Rd, Shanghai 200011, Peoples R China

[4] China Railway First Survey & Design Inst Grp Ltd, 2 Xiying Rd, Xian 710043, Peoples R China

来源：

MEASUREMENT | 2022年 / 190卷

基金：

中国国家自然科学基金;

关键词：

Foreign object detection; Ballastless trackbed; Semisupervised learning; Deep learning; Convolutional neural networks; ANOMALY DETECTION;

D O I：

10.1016/j.measurement.2022.110757

中图分类号：

T [工业技术];

学科分类号：

08 ;

摘要：

This paper proposes a semisupervised algorithm for detecting foreign objects in ballastless beds based on the improved deep SVDD (Support Vector Data Description) algorithm. First, we use the improved Mask R-CNN algorithm to extract the rail and fastener areas in images, assuming that no foreign object exists in the rail and fastener areas. Second, we deepen the backbone network of the deep SVDD to enhance its ability to extract deep semantics from complex images. We perform pure color coverage processing with different colors and mean blur processing with different blur kernels on the rails and fastener regions extracted by the improved Mask R-CNN. The results show that the AUC (Area Under the Curve) of our improved deep SVDD algorithm is 89.23% and improves the AUC compared to that of the original model by 11.09%.

引用

页数：17

共 45 条

[1] GANomaly: Semi-supervised Anomaly Detection via Adversarial Training [J].

Akcay, Samet ;

Atapour-Abarghouei, Amir ;

Breckon, Toby P. .

COMPUTER VISION - ACCV 2018, PT III, 2019, 11363 :622-637

[2]

Alvarez FJ, 2004, 2004 IEEE INTELLIGENT VEHICLES SYMPOSIUM, P866

[3]

Amer M., 2013, P ACM SIGKDD WORKSHO, P8, DOI DOI 10.1145/2500853.2500857

[4] Mask R-CNN [J].

He, Kaiming ;

Gkioxari, Georgia ;

Dollar, Piotr ;

Girshick, Ross .

2017 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2017, :2980-2988

[5]

Aytekin C, 2018, IEEE IJCNN

[6]

Baoqing G., 2012, J SCI INSTRUM, V2

[7]

Cai HL, 2020, IEEE SYS MAN CYBERN, P2776, DOI [10.1109/smc42975.2020.9282881, 10.1109/SMC42975.2020.9282881]

[8] An Intrusion Detection System for the Protection of Railway Assets Using Fiber Bragg Grating Sensors [J].

Catalano, Angelo ;

Bruno, Francesco Antonio ;

Pisco, Marco ;

Cutolo, Antonello ;

Cusano, Andrea .

SENSORS, 2014, 14 (10) :18268-18285

[9]

Chalapathy R., 2018, Anomaly detection using one-class neural networks

[10] Deep Learning for the Detection and Recognition of Rail Defects in Ultrasound B-Scan Images [J].

Chen, Zhengxing ;

Wang, Qihang ;

Yang, Kanghua ;

Yu, Tianle ;

Yao, Jidong ;

Liu, Yong ;

Wang, Ping ;

He, Qing .

TRANSPORTATION RESEARCH RECORD, 2021, 2675 (11) :888-901

← 1 2 3 4 5 →