基于DCNN的井下行人监测方法研究

Research of Underground Pedestrian Monitoring Method Based on DCNN

煤矿安全尤其是井下生产环境的安全一直是煤矿行业的重中之重。大部分煤矿企业对于井下工作人员的检测的智能化水平较低,所采用的人员定位系统大都使用射频卡等技术,无法规避替下、捎卡的情况。系统精准度不高,特别是当监控人员疏忽时,存在很大的安全隐患。基于这样的背景,提出了一种基于DCNN的行人检测技术,针对矿井下的视频质量差、背景单调、检测目标单一等特点对原有的YOLO系统进行了改进,实验结果表明,改进后的YOLO系统对井下特殊环境的检测有比较好的检测效果。

Coal mine safety,especially the safety of underground production environment,has always been the top priority of the coal mining industry.Most of the coal mine enterprises have lower level of intelligence for the detection of underground staff.Most of the personnel location systems used radio frequency cards and other technologies,so they can not avoid the situation of replacing and taking the card.System accuracy is not high,especially when the monitoring staff negligence,there are great security risks.Based on this background,this paper presents a DCNN based pedestrian detection technology,according to the mine under the background of monotonous,poor video quality,the characteristics of single target detection has made the improvement to the existing YOLO system,the experimental results show that the detection of special underground environment improved YOLO system has better detection result.