YOLOv10 and RT-DETR for Weld Defect Detection: Experimental Evaluation
DOI:
https://doi.org/10.58190/ijamec.2026.180Keywords:
CAM, YOLOV10, Deep Learning, object detection, RT-DETR, Weld defect detectionAbstract
Automated weld defect inspection is essential for improving product quality, reducing inspection costs, and minimizing human error in industrial manufacturing. Although deep learning-based object detectors have demonstrated promising performance, comprehensive comparisons under identical experimental conditions remain limited. This study presents a systematic evaluation of YOLOv10 and RT-DETR architectures for weld defect detection using the publicly available Weld Defects (WD) dataset. Five object detection models, namely YOLOv10-N, YOLOv10-L, YOLOv10-X, RT-DETR-L, and RT-DETR-X, were trained and evaluated under identical hyperparameter settings. In addition, the influence of different training–validation split ratios on YOLOv10 performance was investigated. Detection performance was assessed using Precision, Recall, F1-score, mAP@50, and mAP@50–95, while qualitative analyses were conducted through representative detection results and class activation mapping (CAM) visualizations. Experimental results showed that RT-DETR-X achieved the highest localization performance with a mAP@50–95 of 0.458 and the best F1-score, whereas YOLOv10-X obtained the highest precision (0.768) while maintaining competitive overall performance. Explainability analysis further demonstrated that RT-DETR-X generated more spatially coherent attention maps, whereas YOLOv10-X produced more localized but occasionally fragmented activations. The findings provide a comprehensive comparison of CNN- and Transformer-based detectors and offer practical guidance for selecting suitable deep learning models for reliable automated weld defect inspection in industrial applications.
Downloads
References
[1] X. Chen, Y. Duan, Z. Chen, and X. Chen, "Multimodal defect detection in CMT welding based on melt pool images and acoustic signals," in 2025 IEEE International Conference on Robotics and Biomimetics (ROBIO), 2025, pp. 330–335.
[2] G. Chudasama, S. Kashyap, J. Patel, V. Kalyankar, and J. Shewale, "AISI 2205 DSS for body-in-white automotive structure: Effect of welding current and time on the performance of spot welded joint," Welding International, vol. 36, no. 12, pp. 769–778, 2022.
[3] V. Gaidhane and S. Kolhe, "Enhanced weld defect detection through advanced image preprocessing techniques," in 2025 International Conference on Electronics, Computing, Communication and Control Technology (ICECCC), 2025, pp. 1–6.
[4] F. Zuo, J. Liu, M. Fu, L. Wang, and Z. Zhao, "An efficient anchor-free defect detector with dynamic receptive field and task alignment," IEEE Transactions on Industrial Informatics, vol. 20, no. 6, pp. 8536–8547, 2024.
[5] B. Xiao, Y. Yang, Y. He, and G. Ma, "Multi-type weld defect detection in galvanized sheet MIG welding using an improved YOLOv10 model," Materials, vol. 19, no. 6, Art no. 1178, 2026.
[6] C. Zhang, C. Xu, W. Shan, and Z. Han, "DBP-YOLO: A dynamic bifurcated-path fusion network for weld seam defect detection," IEEE Access, vol. 13, pp. 163749–163768, 2025, doi: 10.1109/ACCESS.2025.3609890
[7] K. Parmar, A. Biber, M. Olesch, R. Sharma, and U. Reisgen, "Leveraging thermal images for automatic surface defect detection in TIG welding process of thin metal sheets," Welding in the World, vol. 1, pp. 1–17, 2026.
[8] X. Zhang, H. Chen, Z. Liu, and S. Anli, "Lightweight DETR algorithm for X-ray weld defect detection," Scientific Reports, vol. 15, Art no. 41873, 2025, doi: 10.1038/s41598-025-25916-2.
[9] S. Peng, H. Bao, J. Li, and J. Gao, "YOLO-PDR: A lightweight defect detection network for hot-pressed light guide plates based on parallel texture fusion and multi-scale feature expression," Measurement Science and Technology, vol. 37, no. 12, p. 125902, 2026.
[10] R. Zhang, C. Pan, Z. Chen, J. Fan, Z. Chen, and B. Yuan, "An improved lightweight YOLOv11 algorithm for weld surface defect detection," Scientific Reports, vol. 16, no. 1, Art no. 11440, 2026.
[11] K. Mafuwe, R. V. S. Dulam, C. Kambhamettu, and M. S. Crossley, "Comparative hyperparameter optimization of object detection models for precision monitoring of cucumber beetles and similar insects on yellow sticky cards," Scientific Reports, vol. 16, Art no. 22099, 2026, doi: 10.1038/s41598-026-51483-1.
[12] C. Guo, Y. Du, Y. Yang, S. Wang, D. Liu, and M. Cong, "MCTNet: Defect detection in aluminum alloy die-casting using a multilevel CNN-transformer interaction network," IEEE Transactions on Instrumentation and Measurement, vol. 74, pp. 1–11, 2025, doi: 10.1109/TIM.2025.3589704.
[13] Y. Zhang, H. Zhang, T. Ma, Z. Xu, J. Jiang, and F. Xue, "Weld-YOLO: A YOLO11-based model for high accuracy defect detection in weld X-ray images," in Proc. SPIE, vol. 14247, 2026, Art. no. 1424714.
[14] D. Yang, H. Jiang, H. Yang, Y. Wang, Z. Liu, X. Zhang, and H. Cheng, "Casting defect-detection method based on multi-image feature of fusion artificial intelligence model," Applied Intelligence, vol. 55, no. 11, p. 774, 2025.
[15] W. Zhang, J. Liu, Z. Yan, M. Zhao, X. Fu, and H. Zhu, "FC-YOLO: An aircraft skin defect detection algorithm based on multi-scale collaborative feature fusion," Measurement Science and Technology, vol. 35, no. 11, p. 115405, 2024.
[16] M. Huan, T. Wang, Q. Xu, L. Gao, Z. Lu, X. Shi, M. Qian, L. Jia, and C. Yao, "An efficient lightweight object detection algorithm for defect identification in Paeoniae Radix Alba with rapid area and diameter estimation using OpenCV," Food Science & Nutrition, vol. 13, no. 9, Art. no. e70947, 2025.
[17] A. Dilbaz and İ. A. Ozkan, "Simulation-based spot welding inspection on automotive chassis using YOLO-powered image processing," International Journal of Automotive Engineering and Technologies, vol. 14, no. 3, pp. 170–180, 2025, doi: 10.18245/ijaet.1729908.
[18] A. Dilbaz, "YOLO12 and YOLO13 for steel surface defect detection: Experimental evaluation," Intelligent Methods in Engineering Sciences, vol. 5, no. 2, pp. 55–68, 2026, doi: 10.58190/imiens.2026.175.
[19] A. Wang, H. Chen, L. Liu, K. Chen, Z. Lin, J. Han, and G. Ding, "YOLOv10: Real-time end-to-end object detection," Advances in Neural Information Processing Systems, vol. 37, pp. 107984–108011, 2024, doi: 10.52202/079017-3429.
[20] Roboflow Universe, "Weld defects dataset," 2024. [Online]. Available: https://universe.roboflow.com/yolov8-daldb/weld-defects-mlopr [Accessed: Aug. 1, 2026]
[21] L. Jia, Y. Zhu, and B. Li, "DPCNet: A dual path cross perception network for small object detection in UAV imagery," PLOS ONE, vol. 21, no. 3, Art no. e0344091, 2026.
[22] B. İnan, İ. A. Özkan, and N. Doğan, "Deep learning-based skin disease detection: Comparative performance analysis of YOLOv8 and YOLOv11 models," International Journal of Applied Methods in Electronics and Computers, vol. 13, no. 4, pp. 121–130, 2025.
[23] S. Haitao, Z. Li, X. Xiaoxue, S. Wang, and Y. Zhang, "Rapid identification of mangrove leaves based on improved YOLOv10 model," Forests, vol. 16, no. 7, Art no. 1068, 2025.
[24] Y. Shen and C. Zhang, "Research on the application of YOLOv10n in helmet detection," IAENG International Journal of Computer Science, vol. 53, no. 2, pp. 721–729, 2026.
[25] M. Hussain, "YOLOv5, YOLOv8 and YOLOv10: The go-to detectors for real-time vision," arXiv preprint arXiv:2407.02988, 2024.
[26] L. Liu, R. Yuan, Y. Lv, W. Bo, X. Hu, and E. D. Gedikli, "An efficient high-resolution remote sensing object detection approach using a multilevel optimized RT-DETR model," IEEE Sensors Journal, vol. 25, no. 22, pp. 42306–42315, 2025, doi: 10.1109/JSEN.2025.3617485.
[27] Q. Chen, X. Su, X. Zhang, J. Wang, J. Chen, Y. Shen, C. Han, Z. Chen, W. Xu, F. Li, and S. Zhang, "LW-DETR: A transformer replacement to YOLO for real-time detection," arXiv preprint arXiv:2406.03459, 2024.
[28] Q. Wang, Y. Li, B. Luo, Z. Tu, and C. Li, "Dual-correlation hypergraph network for unaligned RGBT video object detection and a large-scale benchmark," arXiv preprint arXiv:2607.08191, 2026.
[29] F. Liu, F. Wang, M. Liu, and D. Zhang, "RT-DETR-Crack: A tunnel lining crack detection method via multiscale feature enhancement and spatial attenuation mechanism," IAENG International Journal of Applied Mathematics, vol. 56, no. 4, pp. 1593–1606, 2026.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 International Journal of Applied Methods in Electronics and Computers
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
