Improved YOLOv5 for Stroller and Luggage Detection
Jiandang Yang, Qinfeng Tong, Yi Zhong, Qiujian Li · 2023
With economic development and social progress, urban infrastructure is constantly improving. In public places such as shopping malls and subways, escalators are very common and provide great convenience for citizens. However, some irregular behaviours on the escalator have brought great safety hazards. Carrying large luggage and strollers onto the escalator is a typical risky behaviour. How to detect large luggage and strollers in time and issue an alarm is an important part of abnormal detection of escalators. This paper introduces the CBAM module based on YOLOv5, and conducts experiments on the collected data sets. Experimental results show that our proposed model improves the object detection performance well.