基于目标检测的压裂井场全覆盖监控系统  

作　　者：张钊 左威 景江 姜一博 李诚 马涛 陈亮 

机构地区：[1]中国石油集团西部钻探工程有限公司井下作业公司,新疆克拉玛依834000 [2]烟台杰瑞石油装备技术有限公司,山东烟台264000

出　　处：《科技创新与应用》2024年第7期146-149,共4页Technology Innovation and Application

基　　金：中国石油集团西部钻探工程有限公司井下作业公司压裂酸化电子监督系统开发(2022ZC061)。

摘　　要：针对压裂井场作业人员的不安全行为包括作业时闯入高压区、不佩戴安全帽、吸烟、接打电话和不穿着工服作业,使用图像识别目标检测技术,构建井场安全隐患识别算法模型,开发井场全覆盖监控系统。通过采集井场监控画面视频,制作目标检测图像数据集,训练神经网络模型,实现高压区域入侵监测及人的不安全行为提示。采用画面加声音提醒的报警方式,将所有分析结果反馈到仪表橇内。系统使用效果满足当前井场安全监控需求,助力实现井场安全生产目标。In response to the unsafe behaviors of fracturing well site operators,including breaking into high pressure area,not wearing helmet,smoking,answering phone calls,not wearing overalls,using image recognition target detection technology,a well site safety hidden danger identification algorithm model is constructed,and a well site full coverage monitoring system is developed.By collecting the video of the monitoring picture of the well site,making the target detection image data set and training the neural network model,the intrusion monitoring in the high pressure area and the hint of human unsafe behavior are realized.The alarm mode of picture plus sound reminder is adopted to feedback all the analysis results to the instrument sleigh.The use effect of the system meets the needs of current well site safety monitoring and helps to achieve the goal of well site safety production.

关 键 词：压裂井场 目标检测 全覆盖监控系统 高压区域入侵监测 不安全行为提示 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]