ARL-YOLO: A Lightweight Steel Surface Defect Detection Algorithm.

Saved in:

Bibliographic Details
Title:	ARL-YOLO: A Lightweight Steel Surface Defect Detection Algorithm.
Authors:	Sun, Xinguo¹ 410525197@qq.com, Tao, Ye² taibeijack@163.com, Cui, Wenhua³ taibeijack@126.com, Liu, Huiyan⁴ micmoo@126.com, Yu, Yang⁵ 1364744489@qq.com, Sun, Xinli⁶ 501980329@qq.com
Source:	IAENG International Journal of Computer Science. Jul2026, Vol. 53 Issue 7, p2629-2639. 11p.
Subjects:	Algorithms, Object recognition (Computer vision), Edge computing
Abstract:	Steel surface defect detection is the key link of quality control in metallurgical industry, which directly affects product performance and safety production. Steel surface defects are diverse, complex in structure, and belong to the type of small targets. Mainstream detection algorithms suffer from excessive parameters and high computational complexity, hindering their deployment on resource-constrained embedded devices. To address this issue, this paper proposes the ARL-YOLO algorithm based on YOLOv8n, incorporating three key lightweight strategies: (1) replacing the SPPF module with the intra-scale feature interaction module AIFI to enhance small-target feature capture; (2) introducing the RGCSPELAN lightweight module, whose multi-scale fusion mechanism improves perception of objects of varying sizes in complex scenes; (3) designing a lightweight shared convolutional detection head (LSCD) to reduce computational load and parameters while maintaining accuracy. Experimental results show that compared with YOLOv8n, ARL-YOLO achieves a 1.0% accuracy improvement, with model size reduced to 46.9% of the original and parameters down to 1.48MB. This lightweight model lowers deployment costs and is well-suited for edge server applications. [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:	Steel surface defect detection is the key link of quality control in metallurgical industry, which directly affects product performance and safety production. Steel surface defects are diverse, complex in structure, and belong to the type of small targets. Mainstream detection algorithms suffer from excessive parameters and high computational complexity, hindering their deployment on resource-constrained embedded devices. To address this issue, this paper proposes the ARL-YOLO algorithm based on YOLOv8n, incorporating three key lightweight strategies: (1) replacing the SPPF module with the intra-scale feature interaction module AIFI to enhance small-target feature capture; (2) introducing the RGCSPELAN lightweight module, whose multi-scale fusion mechanism improves perception of objects of varying sizes in complex scenes; (3) designing a lightweight shared convolutional detection head (LSCD) to reduce computational load and parameters while maintaining accuracy. Experimental results show that compared with YOLOv8n, ARL-YOLO achieves a 1.0% accuracy improvement, with model size reduced to 46.9% of the original and parameters down to 1.48MB. This lightweight model lowers deployment costs and is well-suited for edge server applications. [ABSTRACT FROM AUTHOR]
ISSN:	1819656X