Research on intelligent safety management of e-commerce logistics storage through a deep learning model

This paper used You Only Look Once version 5 (YOLOv5) and the skeleton extraction algorithm to identify abnormal actions of logistics warehouse personnel. First, the skeleton of personnel actions in consecutive frames of images was obtained using the above two algorithms. Then, a back-propagation neural network (BPNN) was used as the classifier to identify the type of actions. Finally, simulation experiments were conducted. The proposed algorithm achieved strong personnel recognition and localization, skeleton extraction, and action type recognition, while maintaining stable performance when processing multi-person images.