钻井作业典型违章行为视频监控智能分析技术研究

Research on the Typical Violation Intelligent Video Analysis Technology of Drilling Operation

针对钻井现场视频监控人工值守工作量大、存在遗漏、海量视频数据缺乏实时报警等问题,在识别筛选严重事故后果及高概率违章事件的基础上,利用卷积神经网络算法研发视频智能分析技术,开发了钻井视频智能分析与报警系统,研制了配套的便携式视频监控终端及视频分析服务器,对现场直接作业、溢流监测坐岗等过程进行实时行为检测。现场应用表明,该系统可有效识别抽烟、人员离岗等现场典型违章行为,提升了钻井安全管理智能化水平。

For drilling site video monitoring manual duty workload,there are omissions and massive video data lack of real-time alarm,on the basis of the identification and screening of serious accident consequences and high probability of violations,the convolutional neural network algorithm is used to develop video intelligent analysis technology,develop the drilling video intelligent analysis and alarm system,and develop the supporting portable video monitoring terminal and video analysis server,so as to carry out real-time behavior detection in the process of direct field operation and overflow monitoring.It is shown that the system can effectively identify typical violations such as smoking and staff leave,and also improve the level of drilling management intelligence.