### Summary of Research Progress on Application of Prohibited Item Detection in X-Ray Images

LIANG Tianfen, ZHANG Nanfeng, ZHANG Yanxi, YUAN Jinhao, GAO Xiangdong

1. 1.School of Electromechanical Engineering, Guangdong University of Technology, Guangzhou 510006, China
2.Huangpu Customs Technical Center, Dongguan, Guangdong 523076, China
• Online:2021-08-15 Published:2021-08-16

### 违禁品X光图像检测技术应用研究进展综述

1. 1.广东工业大学 机电工程学院，广州 510006
2.黄埔海关技术中心，广东 东莞 523076

Abstract:

X-ray images are widely used in security inspections. At present, most of the security inspections are done manually. However, the heavy workload and work intensity of X-ray security inspections make automatic security inspections an inevitable trend. Therefore, how to automatically detect objects based on X-ray images has become a research hotspot. With the great progress of object detection based on deep learning technology, deep learning models are also widely used in X-ray image prohibited item detection for research and obtain a lot of results. In order to summarize the existing research in a comprehensive and detailed manner, this paper first introduces the characteristics of X-ray images, the traditional methods of X-ray image detection and the methods based on deep learning, then compares the detection effects of traditional methods and deep learning methods, and analyzes the current research progress of automatic security inspection. Finally, in order to provide reference for the research of X-ray image prohibited item detection, this paper points out the research directions worthy of attention in the future.

X光图像在安检中应用十分广泛，目前大部分安检工作还要依靠人工完成，但X光安检巨大的工作量和工作强度使自动安检成为必然趋势。如何根据X光图像自动检测其中物体成为研究热点。随着基于深度学习技术的目标检测取得巨大进展，在X光图像违禁品检测中也大量应用深度学习模型进行研究并获得大量成果。为全面、详细总结现有研究，首先介绍X光成像特点、X光图像检测的传统方法以及基于深度学习的方法，然后对比传统方法与深度学习方法的检测效果并分析现有自动安检研究进展，最后指出未来值得关注的研究方向，以期给X光图像违禁品检测的研究提供参考。