Research on Video Detection of Object Intrusion in Substation
Yi Jiang, Hui Zhi Sun, Hanghang Zhao, Rui Lin, Zhe Li · Journal of Physics Conference Series · 2020
Abstract The safe running of substations is an important part of power system security, as well as an important work of the power companies. The intrusion of kites, plastic trash, small animals and other objects will affect the safe running of substations. With the development of video monitoring technology, intelligent recognition technology of images and videos, video monitoring has gradually become the main monitoring means of unattended substations. This paper firstly analyses the research status and main technologies of video object recognition. To achieve the detection of foreign objects in substations, moving object detection and video object detection technologies are studied. And two outlines are proposed: using background subtraction and image classification, using YOLOv3 to detect objects in videos. In the end of the paper, YOLOv3 was tested by training a detector of small animals. The results show that the value of mAP equals to 0.7169 and the detection speed fulfills the request of real-time detection. This method could serve as reference for detecting and tracking intrusion objects in substations.