The detection of road and cracks is crucial for the safe operation and maintenance of infrastructure, but in practical scenarios, complex backgrounds and multi-scale morphological changes lead to insufficient detection accuracy of YOLOv8n baseline model. In response to this issue, this article introduces the HPA (Hybrid Pooling Attention) module in the YOLOv8n Neck stage, which combines average pooling and max pooling with cross space learning to enhance the transmission and fusion of multi-scale crack features and alleviate information attenuation in feature propagation. The improved model significantly improves the detection rate and positioning accuracy of small cracks in complex backgrounds while maintaining real-time performance, providing an efficient solution for intelligent detection of road and cracks.

Guizhuo Ai, Lan Ma. Research on Road Crack Detection Technology Based on Improved YOLOv8n. Academic Journal of Architecture and Geotechnical Engineering (2026), Vol. 8, Issue 1: 128-133. https://doi.org/10.25236/AJAGE.2026.080116.

[1] Deng Shengbin, Ye Lijia, Xu Na, et al. Research on Road Crack Identification Method Based on Lightweight Algorithm [J]. Highway Engineering, 2025, 50 (05): 150-159.

[2] Xu Zhenkuan, Gao Hang, Li Xuelian, et al. Study on crack propagation behavior of composite base asphalt pavement under dynamic load [J]. Highway Engineering, 2024, 49 (05): 86-92.

[3] Chen Hui, Fu Tao, Yu Yongjie, Zhang Shunxiang. Steel surface defect detection based on HMF-YOLO model [J/OL]. Optoelectronics and Laser, 2023,(05),1‑13.

[4] Fu Yong. Research on Road Crack Detection Based on YOLOv5s Surveying Data Analysis [J]. China New Technology and New Products, 2026, (9): 110-112.

[5] Yang Qing, Liang Yi, Tao Jueqiang. Lightweight pavement disease detection method based on wavelet transform and deformable attention [J]. Journal of Zhejiang Normal University (Natural Science Edition), 2026, 49 (2): 141-151.