| 47 | 0 | 75 |
| 下载次数 | 被引频次 | 阅读次数 |
为确保国家公共安全，有效防范违禁品可能引发的安全隐患，X光安检违禁物品检测技术作为维护公共安全的关键手段之一，其检测准确性与检测效率的提升具有重要意义。对基于深度学习的X光安检违禁物品检测技术进行了探索，重点研究了RT-DETR模型算法，并且在原模型基础上添加RC＿EMA和DySample来改进RT-DETR模型，最后选取OPIXray数据集来进行实验验证。通过消融实验验证了改进后的模型在保持计算效率的同时，精确度、召回率和mAP@0.5指标都有显著提升。指标的提升不仅展示了深度学习在提高公共安全方面的潜力，而且为X光安检违禁物品检测技术的优化提供了有益的参考。
Abstract：To ensure national public safety and effectively prevent security risks caused by prohibited items, X-ray security inspection technology for prohibited items detection, as one of the key measures to maintain public safety, holds significant importance in improving detection accuracy and efficiency. X-ray security inspection based on deep learning has been primarily focused, with special attention given to the RT-DETR algorithm. To improve the performance of RT-DETR model, RC＿EMA and DySample are integrated into the original model. The OPIXray dataset is selected for experimental validation. Ablation experiments verify that the refined model achieves significant improvements in precision, recall rate, and mAP@0.5 metrics while maintaining computational efficiency. These improvements of indicators not only demonstrate the potential of deep learning in enhancing public safety, but also provide valuable references for the optimization of X-ray security inspection technology for prohibited item detection.
[1] 梁添汾，张南峰，张艳喜，等.违禁品X光图像检测技术应用研究进展综述[J].计算机工程与应用，2021,57(16):74-82.
[2] 郝志英，袁得嵛.基于迁移学习的小样本血腥暴力图片识别算法研究[J].中国人民公安大学学报(自然科学版),2024,30(1):58-65.
[3] 陈志强，张丽，金鑫.X射线安全检查技术研究新进展[J].科学通报，2017,62(13):15.
[4] 宋修竹.X光安检图像处理方法研究[D].沈阳：沈阳理工大学，2024.
[5] 张友康，苏志刚，张海刚，等.X光安检图像多尺度违禁品检测[J].信号处理，2020,36(7):1096-1106.
[6] 唐浩漾，王燕，张小媛，等.基于特征金字塔的X光机危险品检测算法[J].西安邮电大学学报，2020,25(2):58-63.
[7] 郭守向，张良.基于单阶段网络的X光图像违禁品检测[J].激光与光电子学进展，2021,58(8):75-84.
[8] REN Y,ZHANG H G,SUN H X,et al.Light Ray:Lightweight network for prohibited items detection in X-ray images during security inspection[J].Computers and Electrical Engineering,2022,103:108.
[9] 李启明，阙祖航.基于改进YOLOv5 的X 射线图像危险品检测[J].Science Technology and Engineering,2023,23(4):1598-1606.
[10] 张娜，罗源，包晓安，等.基于改进Cascade RCNN网络的X光安检违禁品检测[J].计算机系统应用，2022,31(7):224-230.
[11] JIANG C,ZHANG H Y.AM-YOLO:Improved YOLOV4 based on attention mechanism and multi-feature fusion[C]//2022 IEEE 6th Information Technology and Mechatronics Engineering Conference,2022:1403-1407.
[12] 张康佳，张鹏伟，陈景霞，等.基于改进 YOLOv5s 的 X 光图像危险品检测[J].陕西科技大学学报，2023,41(6):176-183+200.
[13] 杨登杰，赵昕，谢悦.基于YOLOv3的X光图像违禁品目标检测[J].工业控制计算机，2024,37(6):50-51.
[14] 庞玉东，李志星，刘伟杰.基于改进实时检测 Transformer 的塔机上俯视场景小目标检测模型[J].计算机应用，2024,44(12):3922-3929.
[15] REN S,HE K,GIRSHICK R,et al.Faster R-CNN:Towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[16] CHEN J,KAO S,HE H,et al.Run,don't walk:Chasing higher FLOPS for faster neural networks[C]//Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition,2023:12021-12031.
[17] SOUDY M,AFIFY Y M,BADR N.RepConv:A novel architecture for image scene classification on intel scenes dataset[J].International Journal of Intelligent Computing and Information Sciences,2022,22(2):63.
[18] ZHU Y,ZHANG Y,ZHANG H.Data augmentation of X-ray images in baggage inspection based on generative adversarial networks[J].IEEE Access,2020,8:86536-86544.
[19] WEI Y,TAO R,WU Z,et al.Occluded prohibited items detection:An X-ray security inspection benchmark and de-occlusion attention module[C]//Proceedings of the 28th ACM International Conference on Multimedia,2020:138-146.
[20] 朱昊，基于深度学习的X光安检违禁品检测算法研究[D].哈尔滨：东北林业大学，2021.
基本信息:
DOI：
中图分类号:X924.4;O434.19;TP18
引用信息:
[1]刘卫华,王蓉,胡配雨等.基于深度学习的X光安检违禁物品检测[J].中国人民公安大学学报(自然科学版),2025,31(02):59-67.
基金信息:
甘肃政法大学科研资助项目(GZF2021XQN15); 甘肃省教育厅高校教师创新基金项目(2023A-100)