城市轨道交通安检中基于X射线图像的危险品识别技术研究
常青青陈嘉敏李维姣
Dangerous Goods Detection Technology Based on X-ray Images in Urban Rail Transit Security Inspection
CHANG QingqingCHEN JiaminLI Weijiao
-
作者信息:公安部第三研究所， 200031， 上海
-
Affiliation:The Third Research Institute of the Ministry of Public Security, 200031, Shanghai, China
-
关键词:
-
Key words:
-
DOI:10.16037/j.1007-869x.2022.04.044
-
中图分类号/CLCN:X924.3∶U231
-
栏目/Col:研究报告
摘要:
阐述了基于传统机器学习和深度学习的危险品对X射线图像安检识别技术原理、方法及适用场景，分析了不同识别技术的优缺点。基于深度学习的危险品识别技术能自动学习物品分类特征，具有良好的鲁棒性和运算效率。其中基于回归思想的目标检测框架的识别速度快，适用于实时系统。利用实际典型场景进行训练并测试，测试结果表明，基于YOLO模型建立的危险物识别技术在识别精度和速度上均能满足相关要求，可对行李包裹内枪支刀具、烟花爆竹等危险品进行智能化识别并报警，能切实提升城市轨道交通安检效能，提高安全风险预警能力。
Abstracts:
Based on X-ray image security inspection equipment, the principle, method and applicable scenarios of dangerous goods detection technology based on conventional machine learning and deep learning are expounded. Merits and demerits of various detection technologies are analyzed. A dangerous goods detection technology based on deep learning can learn the characteristics of goods categorization by itself, having satisfying robustness and calculation efficiency, of which the target-detection framework based on regression ideology can detect fast, applicable to realtime system. Practical typical scenarios are used for training and testing. The test results show that the dangerous goods detection technology built on the basis of YOLO model can meet relevant requirements of identification accuracy and speed, capable of intelligent identification and alarming of dangerous goods such as guns, knives and firecrackers in luggage. The efficiency of urban rail transit security inspection is solidly enhanced, as well as the security risk early warning capability.
引文 / Ref:
常青青, 陈嘉敏, 李维姣. 城市轨道交通安检中基于X射线图像的危险品识别技术研究. 城市轨道交通研究, 2022, 25(4): 205.CHANG Qingqing, CHEN Jiamin, LI Weijiao. Dangerous goods detection technology based on X-ray images in urban rail transit security inspection. Urban Mass Transit, 2022, 25(4): 205.
- 上一篇: 高铁物流运输模式及其可行性
- 下一篇: 城市轨道交通运营安全的第三方评估
过刊浏览
资料下载
友情链接