Improved YOLOV5-Based UAV Pavement Crack Detection

• Published in: IEEE sensors journal Vol. 23; no. 14; pp. 15901 - 15909
• Main Authors: Xing, Jian, Liu, Ying, Zhang, Guang-Zhu
• Format: Journal Article
• Language: English
• Published: New York IEEE 15.07.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Autonomous aerial vehicles; Bidirectional feature pyramid network (BIFPN); Computational modeling; crack; Cracks; Deep learning; Model accuracy; Object recognition; Optimization; Pavements; Pixels; Proposals; Road transportation; Sensors; swin transformer; Transformers; unmanned aerial vehicle (UAV); Unmanned aerial vehicles; YOLOV5
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2023.3281585

In terms of highway crack detection, the combination of unmanned aerial vehicles (UAVs) and deep learning networks has become a powerful detection means. However, in the actual detection, in order to take into account the detection efficiency, it is necessary to ensure that the detection area is large enough, which makes the crack occupy few pixels in the image, and the image background is complex. Therefore, in this article, DJI Mavic3 is used to establish the image data set of highway pavement cracks under complex background. And, the YOLOV5 deep learning model is improved by adding a swin transformer structure and bidirectional feature pyramid network (BIFPN) feature pyramid. The improved YOLOV5 model achieved real-time pixel-level detection with a detection accuracy of 90% and a detection speed of 43.5 FPS. In terms of crack detection ability, the accuracy of the improved YOLOV5 reaches four pixels, and cracks of 1.2 mm can be detected in the experiment. Compared with the YOLOV7 model, the detection accuracy of the improved YOLOV5 model is increased by 15.4%. Compared with the YOLOV6 model, the calculated parameters of the improved YOLOV5 model are reduced by 59.25%. The proposed model is superior to other advanced models in crack detection.