摘要
为了保障矿工安全、保护设备及提高生产效率,提出基于多传感器信息融合的露天煤矿大型工程车辆识别与闯入预警技术。首先,在整体技术架构和实现方法设计的基础上,提出基于YOLOv8的工程车辆识别与检测方法;然后,在露天煤矿现场搭建检测软硬件平台,通过现场拍摄、网络搜集共计6300张样本数据集,测试基于YOLOv8的工程车辆识别与检测方法的识别精度。结果表明:文中所提出的基于YOLOv8工程车辆检测模型,可以快速准确地识别露天煤矿作业现场内出现的多个车辆目标,检测准确率在0.85以上,漏检率较低,而且,能够识别不完整的车辆图像。提出的工程车辆识别与闯入预警系统,可为工作中的设备提供来车识别与预警提示,有助于预防安全事故的发生。
To ensure the safety of miners,protect equipment,and improve production efficiency,the identification and intrusion early warning technology of large-scale engineering vehicles in opencast coal mines based on multi-sensor information fusion was studied.Firstly,based on the overall technical framework and implementation method,the identification and detection method of engineering vehicles based on YOLOv8 was proposed.The software and hardware platform for detection was built in the opencast coal mine,and the identification accuracy of the engineering vehicle identification and detection method based on YOLOv8 was tested through a total of 6300 sample datasets from on-site shooting and networks.The results show that the engineering vehicle detection model based on YOLOv8 can quickly and accurately identify multiple vehicle targets in the opencast coal mine,and the detection accuracy is more than 0.85,with a low missed detection rate.In addition,the incomplete vehicle image can be recognized.The engineering vehicle identification and intrusion early warning system studied in this paper provides vehicle identification and early warning hints for the working equipment to avoid safety accidents.
作者
孙国思
马鹏飞
王威淳
郜普浩
朱健伟
SUN Guosi;MA Pengfei;WANG Weichun;GAO Puhao;ZHU Jianwei(Open-pit Coal Mine of Baori Shiller Energy Co.,Ltd.,National Energy Group,Hulunbuir Inner Mongolia 021008,China)
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2024年第S01期191-198,共8页
China Safety Science Journal
关键词
YOLOv8
露天煤矿
大型工程车辆
目标识别
闯入安全预警
YOLOv8
opencast coal mine
large-scale engineering vehicle
target identification
intrusion early warning
