广西师范大学学报（自然科学版） ›› 2021, Vol. 39 ›› Issue (4): 34-46.doi: 10.16088/j.issn.1001-6600.2020080902

• • 上一篇    下一篇

基于RFB网络的特征融合管制物品检测算法研究

张伟彬, 吴军, 易见兵*   

  1. 江西理工大学 信息工程学院, 江西 赣州 341000
  • 收稿日期:2020-08-09 修回日期:2020-12-08 出版日期:2021-07-25 发布日期:2021-07-23
  • 通讯作者: 易见兵(1980—), 男, 江西宜春人, 江西理工大学讲师, 博士。E-mail: yijianbing8@163.com
  • 基金资助:
    国家自然科学基金(61862031); 江西省自然科学基金(20181BAB202004); 江西省教育厅科技项目(GJJ190458)

Research on Feature Fusion Controlled Items Detection Algorithm Based on RFB Network

ZHANG Weibin, WU Jun, YI Jianbing*   

  1. College of Information Engineering, Jiangxi University of Science and Technology, Ganzhou Jiangxi 341000, China
  • Received:2020-08-09 Revised:2020-12-08 Online:2021-07-25 Published:2021-07-23

PDF (PC)

185

摘要： 为了提高对管制物品的检测精度,本文提出一种结合RFB(receptive field block)网络结构和特征融合的目标检测算法。首先对采集的安检数据进行无效内容剔除、滤波;接着对安检数据进行人工标注和数据增强;然后在MobileNetV3-SSD算法的基础上,通过引入RFB网络改进其网络结构,以加强网络的特征提取能力,并利用特征融合的方法提高模型的小目标检测能力;最后,构建了一个安检数据集SCCI2020来验证算法的性能,该数据集包含91 767张图片。实验结果表明,本算法在安检数据集SCCI2020上的检测精度为87.0%,比MobileNetV3-SSD算法的检测精度高2.7个百分点;在COCO2014和COCO2017通用数据集上的检测精度分别为21.9%和23%,相对于VGG16-SSD、MobileNetV3-SSD算法均有一定提升。

关键词: 管制物品检测, 小目标, RFB网络, 特征融合, MobileNetV3-SSD

Abstract: In order to improve the detection accuracy of controlled items, a target detection algorithm that combines Receptive Field Block (RFB) network structure and feature fusion is proposed. First, the collected security inspection data are eliminated and filtered, then, the security inspection data are manually labeled and the data are enhanced. Based on the MobileNetV3-SSD algorithm, the RFB network is introduced to improve its network structure to strengthen the network’s feature extraction capabilities, and the feature fusion method is used to improve the model’s small target detection capability. Finally, a security inspection dataset SCCI2020 is constructed to verify the performance of the algorithm, where the dataset contains 91 767 images. Experimental results show that the detection accuracy of this algorithm on the security inspection dataset SCCI2020 is 87.0%, which is 2.7 percentage point higher than the detection accuracy of the MobileNetV3-SSD algorithm. The detection accuracy on the COCO2014 general dataset and COCO2017 general dataset are 21.9% and 23%. Compared with VGG16-SSD and MobileNetV3-SSD algorithm, the detection accuracy is improved to a certain extent.

Key words: controlled items detection, small target, RFB network, feature fusion, MobileNetV3-SSD

中图分类号: 

  • TP391.41
[1]JIAO L C, ZHAO J. A survey on the new generation of deep learning in image processing[J]. IEEE Access, 2019, 7: 172231-172263. DOI:10.1109/ACCESS.2019.2956508.
[2]ZHANG S Q, PAN X Z, CUI Y L, et al. Learning affective video features for facial expression recognition via hybrid deep learning[J]. IEEE Access, 2019, 7: 32297-32304. DOI:10.1109/ACCESS.2019.2901521.
[3]WANG K D, LI S Y, NIU S S, et al. Detection of infrared small targets using feature fusion convolutional network[J]. IEEE Access, 2019, 7: 146081-146092. DOI:10.1109/ACCESS.2019.2944661.
[4]LIN L K, WANG S Y, TANG Z X. Using deep learning to detect small targets in infrared oversampling images[J]. Journal of Systems Engineering and Electronics, 2018, 29(5): 947-952. DOI:10.21629/JSEE.2018.05.07.
[5]GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]// 2014 IEEE Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2014: 580-587. DOI:10.1109/CVPR.2014.81.
[6]HE K M, ZHANG X Y, REN S Q, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9): 1904-1916. DOI:10.1109/TPAMI.2015.2389824.
[7]GIRSHICK R. Fast R-CNN[C]// 2015 IEEE International Conference on Computer Vision. Los Alamitos, CA: IEEE Computer Society, 2015: 1440-1448. DOI:10.1109/ICCV.2015.169.
[8]REN S Q, HE K M, 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. DOI:10.1109/TPAMI.2016.2577031.
[9]REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2016: 779-788. DOI:10.1109/CVPR.2016.91.
[10]REDMON J, FARHADI A. YOLOv3: an incremental improvement[EB/OL].(2018-04-08)[2020-07-10]. https://arxiv.org/abs/1804.02767v1.
[11]LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]// Computer Vision: ECCV 2016. Berlin: Springer, 2016: 21-37. DOI:10.1007/978-3-319-46448-0_2.
[12]HOWARD A G, ZHU M L, CHEN B, et al. MobileNets: efficient convolutional neural networks for mobile vision applications[EB/OL].(2017-04-17)[2020-07-10].https://arxiv.org/abs/1704.04861v1.
[13]SANDLER M, HOWARD A G, ZHU M L, et al. MobileNetV2: inverted residuals and linear bottlenecks[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2018: 4510-4520. DOI:10.1109/CVPR.2018.00474.
[14]LI Z X, ZHOU F Q. FSSD: feature fusion single shot mulitibox detector[EB/OL].(2018-05-17)[2020-7-10]. https://arxiv.org/abs/1712.00960v3.
[15]CAO G M, XIE X M, YANG W Z, et al. Feature-fused SSD: fast detection for small objects[C]// Proceedings of SPIE Volume 10615: Ninth International Conference on Graphic and Image Processing (ICGIP 2017). Bellingham, WA: SPIE, 2018: 106151E. DOI:10.1117/12.2304811.
[16]LIU S T, HUANG D, WANG Y H. Receptive field block net for accurate and fast object detection[C]// Computer Vision: ECCV 2018. Berlin: Springer, 2018: 404-419. DOI:10.1007/978-3-030-01252-6_24.
[17]HOWARD A, SANDLER M, CHEN B, et al. Searching for MobileNetV3[C]// 2019 IEEE/CVF International Conference on Computer Vision. Los Alamitos, CA: IEEE Computer Society, 2019: 1314-1324. DOI:10.1109/ICCV.2019.00140.
[18]MIAO C J, XIE L X, WAN F, et al. SIXray: a large-scale security inspection X-ray benchmark for prohibited item discovery in overlapping images[C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2019: 2114-2123. DOI:10.1109/CVPR.2019.00222.
[19]DAI J F, LI Y, HE K M, et al. R-FCN: object detection via region-based fully convolutional networks[EB/OL].(2016-05-20)[2020-7-10]. https://arxiv.org/abs/1605.06409.
[20]LIN T Y, DOLLÁR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2017: 936-944. DOI:10.1109/CVPR.2017.106.
[21]REN J, CHEN X H, LIU J B, et al. Accurate single stage detector using recurrent rolling convolution[EB/OL].(2017-04-19)[2020-7-10]. https://arxiv.org/abs/1704.05776.
[22]SZEGEDY C, IOFFE S, VANHOUCKE V, et al. Inception-v4, inception-ResNet and the impact of residual connections on learning[C]// AAAI'17: Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence. Palo Alto, CA: AAAI Press, 2017: 4278-4284. DOI:10.5555/3298023.3298188.
[23]王俊强, 李建胜, 周学文, 等. 改进的SSD算法及其对遥感影像小目标检测性能的分析[J]. 光学学报, 2019, 39(6): 0628005. DOI:10.3788/AOS201939.0628005.
[24]雷霆,谢榕昌, 黄滔, 等. 基于SSD改进算法的电缆隧道积水识别方法[J]. 广东电力, 2019, 32(9): 131-136.
[25]任宇杰, 杨剑, 刘方涛, 等. 基于SSD和MobileNet网络的目标检测方法的研究[J]. 计算机科学与探索, 2019, 13(11): 1881-1893. DOI:10.3778/j.issn.1673-9418.1906023.
[26]吉祥凌, 吴军, 易见兵, 等. 基于深度学习的管制物品自动检测算法研究[J]. 激光与光电子学进展, 2019, 56(18): 180402.
[1] 白捷, 高海力, 王永众, 杨来邦, 项晓航, 楼雄伟. 基于多路特征融合的Faster R-CNN与迁移学习的学生课堂行为检测[J]. 广西师范大学学报（自然科学版）, 2020, 38(5): 1-11.
[2] 张灿龙, 李燕茹, 李志欣, 王智文. 基于核相关滤波与特征融合的分块跟踪算法[J]. 广西师范大学学报（自然科学版）, 2020, 38(5): 12-23.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
[1] 戴云飞, 祝龙记. 应用于超级电容储能的开关准Z源双向DC/DC变换器研究[J]. 广西师范大学学报（自然科学版）, 2021, 39(3): 11 -19 .
[2] 吕惠炼, 胡维平. 基于端到端深度神经网络的语音情感识别研究[J]. 广西师范大学学报（自然科学版）, 2021, 39(3): 20 -26 .
[3] 胡锦铭, 韦笃取. 不同阶次分数阶永磁同步电机的混合投影同步[J]. 广西师范大学学报（自然科学版）, 2021, 39(4): 1 -8 .
[4] 武康康, 周鹏, 陆叶, 蒋丹, 闫江鸿, 钱正成, 龚闯. 基于小批量梯度下降法的FIR滤波器[J]. 广西师范大学学报（自然科学版）, 2021, 39(4): 9 -20 .
[5] 刘东, 周莉, 郑晓亮. 基于SA-DBN的超短期电力负荷预测[J]. 广西师范大学学报（自然科学版）, 2021, 39(4): 21 -33 .
[6] 王金艳, 胡春, 高健. 一种面向知识编译的OBDD构造方法[J]. 广西师范大学学报（自然科学版）, 2021, 39(4): 47 -54 .
[7] 逯苗, 何登旭, 曲良东. 非线性参数的精英学习灰狼优化算法[J]. 广西师范大学学报（自然科学版）, 2021, 39(4): 55 -67 .
[8] 李莉丽, 张兴发, 李元, 邓春亮. 基于高频数据的日频GARCH模型估计[J]. 广西师范大学学报（自然科学版）, 2021, 39(4): 68 -78 .
[9] 李松涛, 李群宏, 张文. 三自由度碰撞振动系统的余维二擦边分岔与混沌控制[J]. 广西师范大学学报（自然科学版）, 2021, 39(4): 79 -92 .
[10] 赵红涛, 刘志伟. λ重完全二部3-一致超图λK(3)n,n分解为超图双三角锥[J]. 广西师范大学学报（自然科学版）, 2021, 39(4): 93 -98 .
版权所有 © 广西师范大学学报（自然科学版）编辑部
地址：广西桂林市三里店育才路15号　邮编：541004
电话：0773-5857325　E-mail: gxsdzkb@mailbox.gxnu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发