This paper mainly aims at the problems existing in the safety maintenance of overhead transmission lines, including the poor stability and low adaptability caused by the insufficient locking adaptability of the dwell device to different sizes of ground wires. In addition, during live working on transmission lines, there is also a lack of safety awareness among operators. The problems with these safety violations include not wearing helmets, safety belts and safety harnesses. To this end, this paper proposes a design of an overhead transmission line safety violation identification and stationing device based on a stationing drone and YOLOv5 algorithm. The scheme aims to improve the efficiency and accuracy of safety maintenance, and provide a strong guarantee for the safe operation of overhead transmission lines. Firstly, this paper designs the dwell device to improve its locking adaptability to different sizes of ground wires. After that, collect the image and video data of overhead transmission lines, and then annotate them for data preprocessing and preparation. Finally, the safety helmet, safety belt, safety harnesses and red vest detection model based on YOLOv5 algorithm is used to identify whether the operators are wearing safety helmet, safety belt and safety rope, and whether there are personnel wearing red vest at the work site. Furthermore, a safety violation behavior recognition algorithm framework based on human key point detection and attention mechanism has been proposed to effectively identify four types of safety violations, including not wearing a helmet, not wearing a seat belt, not tying a safety rope, and whether wearing a red vest. The experimental results show that the YOLOv5 algorithm based safety violation behavior recognition model has high accuracy and recall rate, effectively identifying safety violations in overhead transmission lines. This study provides a new intelligent solution for the safe maintenance of overhead transmission lines, which has certain practical application value.
|