A Two-Stage Automatic Framework For Rail Surface Inspection Using YOLOv10 And AutoML.

Saved in:

Bibliographic Details
Title:	A Two-Stage Automatic Framework For Rail Surface Inspection Using YOLOv10 And AutoML.
Authors:	Ha, Minh Tuan¹ minhtuanha031@gmail.com, Nguyen, Phuong Ty² tynp2109@gmail.com
Source:	IAENG International Journal of Computer Science. Jul2026, Vol. 53 Issue 7, p2764-2777. 14p.
Subjects:	Inspection & review, Object recognition (Computer vision), Neural architecture search, Railroad safety measures, Real-time computing, Convolutional neural networks
Abstract:	Automated visual inspection of rail surfaces is critical for railway safety but faces challenges such as complex backgrounds, diverse defects, and environmental noise. This paper proposes a fully automated, two-stage framework integrating YOLOv10 and Automated Machine Learning (AutoML). First, YOLOv10 detects and extracts the rail region to eliminate background interference. Second, an AutoML pipeline automatically designs an optimal hybrid convolutional neural network (CNN) for defect classification, combining components from VGG16, MobileNet, and EfficientNetV2. Experimental results demonstrate the framework's high performance, achieving a 99.3% classification accuracy on an independent test set while maintaining a processing speed suitable for real-time applications. This work highlights the efficacy of combining targeted region extraction with AutoML for robust and efficient rail surface defect inspection. [ABSTRACT FROM AUTHOR]
	Copyright of IAENG International Journal of Computer Science is the property of International Association of Engineers (IAENG) and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database:	Engineering Source

Description
Abstract:	Automated visual inspection of rail surfaces is critical for railway safety but faces challenges such as complex backgrounds, diverse defects, and environmental noise. This paper proposes a fully automated, two-stage framework integrating YOLOv10 and Automated Machine Learning (AutoML). First, YOLOv10 detects and extracts the rail region to eliminate background interference. Second, an AutoML pipeline automatically designs an optimal hybrid convolutional neural network (CNN) for defect classification, combining components from VGG16, MobileNet, and EfficientNetV2. Experimental results demonstrate the framework's high performance, achieving a 99.3% classification accuracy on an independent test set while maintaining a processing speed suitable for real-time applications. This work highlights the efficacy of combining targeted region extraction with AutoML for robust and efficient rail surface defect inspection. [ABSTRACT FROM AUTHOR]
ISSN:	1819656X