Yolo-C:基于单阶段网络的X光图像违禁品检测  

Yolo-C:One-Stage Network for Prohibited Items Detection Within X-Ray Images

作　　者：郭守向 张良[1] Guo Shouxiang;Zhang Liang(Tianjin Key Laboratory of Advanced Signal&Image Processing,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学天津市智能信号与图像处理重点实验室,天津300300

出　　处：《激光与光电子学进展》2021年第8期67-76,共10页Laser & Optoelectronics Progress

基　　金：国家自然科学基金(61179045);民航安全能力建设项目(20600523)。

摘　　要：为了检测X光图像中的违禁物品,提出一种基于深度学习的单阶段双网络目标检测算法。在单阶段目标检测网络Yolov3的基础上,结合复合骨干网络的思想,构建了Yolo-C目标检测网络。Yolo-C的骨干网DarkNet-C由辅助骨干网络DarkNet-A和引导骨干网络DarkNet-L组成。DarkNet-A中的各个特征层与DarkNet-L中对应的上一层级进行特征级联,然后向下一层级传播,最终得到表征图像信息的特征图。为提升对小目标的检测性能,引入特征增强模块(FAB)。对级联后的特征图进行特征融合,以增强特征的非线性表达能力,达到特征平滑的目的。此外,采用迁移学习和数据增强的方法训练网络,提升了网络的鲁棒性。该算法在SIXray_OD数据集上平均精度均值(mAP)达到了73.68%,检测速度达40frame·s~(-1)。实验结果表明,Yolo-C在检测X光图像领域,有效提高了对多类违禁物品的检测精度,且满足检测的实时性要求。To detect prohibited items in X-ray images,this study proposed a one-stage dual-network object detection algorithm based on deep learning.Based on the one-stage object detection algorithm Yolov3 and combined with the idea of a composite backbone network,a Yolo-C object detection network is developed.The backbone of Yolo-C(DarkNet-C)consists of an assistant backbone network(Darknet-A)and a lead backbone network(Darknet-L).Each feature layer of the DarkNet-A is cascaded by feature with the upper feature level corresponding to DarkNet-L and then propagated to the next feature level.Finally,a feature map representing image information is obtained.The feature enhancement block(FAB)is introduced to improve detection performance of small object.Feature fusion is performed on the cascaded feature maps to enhance the nonlinear expression ability of features and achieve the purpose of feature smoothing.Besides,transfer learning and data enhancement was adopted to train the network and improve its robustness.The mAP in the SIXrayOD dataset is 73.68%,and detection speed is 40 frame·s-1.In the X-ray image detection field,Yolo-C has effectively improved the detection accuracy of different prohibited items and met the real-time requirements of detection.

关 键 词：图像处理 违禁品检测 单阶段双网络 Yolo-C 特征增强模块 迁移学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]