Citation

Yang D, Cao Y, Feng Y, Lai X, Pan Z. IEEE Int. Conf. Unmann. Syst. (ICUS Proc.) 2021; 512-519.

Copyright

(Copyright © 2021, IEEE (Institute of Electrical and Electronics Engineers))

DOI

10.1109/ICUS52573.2021.9641291

PMID

unavailable

Abstract

Intelligent lifebuoy is a new type of unmanned high-speed surface rescue robot, but at present it mainly relies on remote control. An intelligent lifebuoy with the ability of drownning detection, distance detection and autonomous navigation will have a good application prospect in the field of surface rescue. This paper analyzes the characteristics of drowning person and the requirement of rescue control and proposes a drowning detection algorithm for intelligent lifebuoy. An improved YOLOv4 network is designed to detect the drowning person and a geometric distance measurement method based on the bounding box is designed to detect the position. The experimental results show that the algorithm can not only achieve a high detection accuracy, but also get the distance and direction synchronously, so as to help the intelligent lifebuoy to complete a fast and effective rescue.

Keywords

Conferences; Water; Machine vision; Clothing; Detection algorithms; distance measurement; Distance measurement; drownning; intelligent lifebuoy; neural network; object detection; Path planning; surface rescue