Article Details

Abstract

With the development of PCBs toward higher density, complexity and miniaturization, higher demands are placed on small target defect detection under complex backgrounds. To improve the model's ability to detect small target defects, this study proposes an enhanced PCB small target defect detection algorithm, YOLO-PMAA, based on YOLOv11. First, a pinwheel convolution (PConv) is introduced into the shallow layers of the backbone network to enlarge the receptive field through asymmetric padding and convolution, ensuring that the features of small targets are transmitted to the deeper layers of the network. Second, a multidimensional cascade hybrid attention module (MDCA) is designed, which focuses on important information across the scale, channel and spatial dimensions, thereby strengthening the feature extraction of thin, long and narrow small targets. Finally, an adaptive shape loss function (ASIoU) is designed to improve the detection performance of the model. The algorithm is validated over multiple rounds on the PKU-Market-PCB dataset. Compared with YOLOv11n, it improves precision by 3.0%, recall by 6.2%, mAP@0.5 by 3.8% and mAP@0.5:0.95 by 3.2%, and reduces the number of parameters by 4.25% and the weight file size by 3.82%, effectively improving model performance while balancing model complexity.

Keywords

PCB defect detection; attention mechanism; loss function; feature fusion; small target detection

References

1. [1] Ling Q，Isa N A M. Printed Circuit Board Defect Detection Methods Based on Image Processing, Machine Learning and Deep Learning: A Survey [J].IEEE Access，2023（11）：15921-15944.
2. [2] Xia K W，Lyu Z L，Liu K，et al. Global Contextual Attention Augmented YOLO with ConvMixer Prediction Heads for PCB Surface Defect Detection [J].Scientific Reports，2023，13（1）：9805-9805.
3. [3] Li B，Palayew S，Li F，et al. PCBDet: An Efficient Deep Neural Network Object Detection Architecture for Automatic PCB Component Detection on the Edge [J/OL].Electronics Letters，2024，60（2）：e13042.https://doi.org/10.1049/ell2.13042.
4. [4] LI D，BAI F，LI S，et al. EL-PCBNet: An Efficient and Lightweight Network for PCB Defect Detection [J/OL].Measurement，2025，253：11771910.1016/j.measurement.2025.117719.
5. [5] Zheng H，Peng J，Yu X，et al. FDDC-YOLO: an efficient detection algorithm for dense small-target solder joint defects in PCB inspection [J/OL].Journal of Real-Time Image Processing，2025，22（2）：83.https://link.springer.com/article/10.1007/s11554-025-01664-4.
6. [6] Xiao G S，Hou S L，Zhou H Y. PCB Defect Detection Algorithm Based on CDI-YOLO [J/OL].Scientific Reports，2024，14（1）：https://doi.org/10.1038/s41598-024-57491-3.
7. [7] Wei Jiamin, Xu Xinying, Han Hui, etc Lightweight PCB defect detection algorithm based on YOLO HorL [J]. Computer Engineering and Design, 2025, 46 (11): 3283-3291
8. [8] Yang J N，Liu S L，Wu J J，et al. Pinwheel-shaped Convolution and Scale-based Dynamic Loss for Infrared Small Target Detection [C]//Proceedings of the AAAI Conference on Artificial Intelligence，2025，39（9）：9202-9210.
9. [9] Zhang Y，Zhang H，Zhao H M. DsP-YOLO: An anchor-free network with DsPAN for small object detection of multiscale defects [J/OL].Expert Systems with Applications，2024，241：122669.https://doi.org/10.1016/j.eswa.2023.122669.
10. [10] Zhu H Y，Wei L Y，Li H. VR-YOLO: Enhancing PCB Defect Detection with Viewpoint Robustness Based on YOLO [C]//2025 IEEE 11th World Forum on Internet of Things (WF-IoT)，2025：1-6.

Funding

Scientific Research Project of Hunan Provincial Department of Education (25C0994)

About the authors

Zhou Yangmiao (1991-), female, Hui ethnicity, from Wuhan, Hubei, lecturer, master's degree, research interests include computer vision.