An integrated application platform for safety management and control of smart power plants based on video image technology
Wenhua Lin, Jianchao Xue, Chuanfang Zhu, Fan Jiang, Cong Ding · 2022 IEEE 10th Joint International Information Technology and Artificial Intelligence Conference (ITAIC) · 2022
In view of the complex environment of the power plant and the difficulty of identifying personnel violations in time, this paper proposes a personnel violation identification technology based on video recognition technology, and builds a relevant platform for engineering verification. This paper first builds the YOLOv5 model for video recognition and extraction, and then uses the HRNet network to extract personal model information for the complex power plant environment. For common violations, the algorithm is further used to extract the human body area to improve the recognition efficiency and accuracy. The article builds a relevant algorithm software platform. Engineering practice shows that the algorithm and platform shown in the article can effectively identify illegal behaviors and have high engineering practice value.