Computational Technology Resources - CCC

J. Song¹, K. Xue¹, K.C. Fung^1,2, K.H. Lin¹, V.T.Y. Ng¹ and K.M. Lam^1,2

¹Centre for Advances in Reliability and Safety, AIR@InnoHK Research Cluster, Hong Kong, China
²Department of Electrical and Electronic Engineering, The Hong Kong Polytechnic University, Hong Kong, China

doi:10.4203/ccc.7.4.7

J. Song, K. Xue, K.C. Fung, K.H. Lin, V.T.Y. Ng, K.M. Lam, "Computer Vision-Based Anomaly Detection on Pantograph Carbon Strip", in J. Pombo, (Editor), "Proceedings of the Sixth International Conference on Railway Technology: Research, Development and Maintenance", Civil-Comp Press, Edinburgh, UK, Online volume: CCC 7, Paper 4.7, 2024, doi:10.4203/ccc.7.4.7

Keywords: anomaly detection, pantograph-catenary system, generative model, image processing, model ensemble, computer vision.

Abstract

This paper presents a computer vision-based AI system designed for detecting defects on the surface of carbon strips. Recognizing the limitations arising from inadequate representation of defective classes, we proposed a data augmentation approach that combines generative models and image processing, incorporating a semi-automated image selection process. Additionally, we have adopted a model ensemble technique to enhance identification accuracy. Through experimental validation, we demonstrated the effectiveness of our data augmentation methodology and model ensemble, resulting in an improved defect recall rate of 91.7%, false alarm recall rate of 93.9%, and accuracy of 92.9%.

download the full-text of this paper (PDF, 8 pages, 738 Kb)

go to the previous paper
go to the next paper
return to the table of contents
return to the volume description

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCC CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Conferences ISSN 2753-3239 CCC: 7 PROCEEDINGS OF THE SIXTH INTERNATIONAL CONFERENCE ON RAILWAY TECHNOLOGY: RESEARCH, DEVELOPMENT AND MAINTENANCE Edited by: J. Pombo Paper 4.7 Computer Vision-Based Anomaly Detection on Pantograph Carbon Strip J. Song¹, K. Xue¹, K.C. Fung^1,2, K.H. Lin¹, V.T.Y. Ng¹ and K.M. Lam^1,2 ¹Centre for Advances in Reliability and Safety, AIR@InnoHK Research Cluster, Hong Kong, China ²Department of Electrical and Electronic Engineering, The Hong Kong Polytechnic University, Hong Kong, China doi:10.4203/ccc.7.4.7 Full Bibliographic Reference for this paper J. Song, K. Xue, K.C. Fung, K.H. Lin, V.T.Y. Ng, K.M. Lam, "Computer Vision-Based Anomaly Detection on Pantograph Carbon Strip", in J. Pombo, (Editor), "Proceedings of the Sixth International Conference on Railway Technology: Research, Development and Maintenance", Civil-Comp Press, Edinburgh, UK, Online volume: CCC 7, Paper 4.7, 2024, doi:10.4203/ccc.7.4.7 Keywords: anomaly detection, pantograph-catenary system, generative model, image processing, model ensemble, computer vision. Abstract This paper presents a computer vision-based AI system designed for detecting defects on the surface of carbon strips. Recognizing the limitations arising from inadequate representation of defective classes, we proposed a data augmentation approach that combines generative models and image processing, incorporating a semi-automated image selection process. Additionally, we have adopted a model ensemble technique to enhance identification accuracy. Through experimental validation, we demonstrated the effectiveness of our data augmentation methodology and model ensemble, resulting in an improved defect recall rate of 91.7%, false alarm recall rate of 93.9%, and accuracy of 92.9%. download the full-text of this paper (PDF, 8 pages, 738 Kb) go to the previous paper go to the next paper return to the table of contents return to the volume description
Back to top	©Civil-Comp Limited 2023 - terms & conditions