Recognition of Unsafe Onboard Mooring and Unmooring Operation Behavior Based on Improved YOLO-v4 Algorithm

In the maritime industry, unsafe behaviors exhibited by crew members are a significant factor contributing to shipping and occupational accidents. Among these behaviors, operation of mooring lines is particularly prone causing severe Video-based monitoring has been demonstrated as an effective means detecting in real time providing early warning members. To this end, paper presents dataset comprising videos line operations on M.V. YuKun. Additionally, we propose behavior recognition model based improved You Only Look Once (YOLO)-v4 network. Experimental results indicate that proposed model, when compared other models such original YOLO-v4 YOLO-v3, demonstrates improvement speed approximately 35% while maintaining accuracy. it also reduction computation burden. Furthermore, was successfully applied actual ship test, which further verifies its effectiveness recognizing behaviors. Results test highlight model’s accuracy par with network but shows processing 50% complexity about 96%. Hence, work can effectively detect potentially enhance safety marine operations.