基于改进SSD算法的车辆检测

doi:10.19678/j.issn.1000-3428.0060031

摘要/Abstract

摘要： SSD算法利用多尺度特征图进行分类和位置回归，检测小目标效果优于YOLO算法，但SSD算法在进行车辆检测时存在漏检问题。为此，提出一种改进SSD算法。为提取更多的车辆特征信息，设计改进Inception模块替代SSD网络中的Conv8、Conv9和Conv10层。将浅层特征的位置信息和深层特征的语义信息进行均衡化融合，构建多尺度特征融合均衡化网络，提高小目标车辆识别率。在特征提取层均引入SENet，对不同特征通道的重要性进行重标定以提高模型性能。实验结果表明，改进后SSD算法在自制的车辆数据集上平均精度为90.89%，检测速度达到59.42 frame/s，相比改进前的SSD算法，在精度和速度上分别提高2.65个百分点和17.41 frame/s，能够更快速、准确地对图像中的车辆进行识别和定位。

关键词: 车辆检测, SSD算法, Inception结构, 注意力机制, 特征融合

Abstract: The Single Shot Multibox Detector(SSD) algorithm uses multi-scale feature maps for classification and position regression.This algorithm displays better performance than the YOLO algorithm in detecting small targets, but frequently misses targets in vehicle detection.To address the problem, this paper proposes an improved SSD algorithm. In order to extract more vehicle feature information, an improved Inception module is given to replace the Conv8, Conv9 and Conv10 layers in the SSD network.At the same time, a network is designed to balance and fuse multi-scale features.The network can fuse the location information of shallow features and the semantic information of deep features, so the accuracy of small target detection can be increased.To further improve model performance, SENet is introduced into feature extraction layers to recalibrate the importance of different feature channels.The experimental results show that the improved SSD algorithm displays an average accuracy of 90.89% and a detection speed of 59.42 frame/s on the self-made vehicle dataset.Compared with the original SSD algorithm, the improved SSD algorithm increases the accuracy by 2.65 percentage points and speed by 17.41 frame/s.This algorithm can identify and locate the vehicles in images more quickly and accurately.

Key words: vehicle detection, SSD algorithm, Inception structure, attentional mechanism, feature fusion

中图分类号:

TP391

李国进, 胡洁, 艾矫燕. 基于改进SSD算法的车辆检测[J]. 计算机工程, 2022, 48(1): 266-274.

LI Guojin, HU Jie, AI Jiaoyan. Vehicle Detection Based on Improved SSD Algorithm[J]. Computer Engineering, 2022, 48(1): 266-274.

http://www.ecice06.com/CN/Y2022/V48/I1/266

图/表 12

20220108123822

20220108123826

20220108123830

20220108123834

20220108123837

20220108123841

20220108123845

20220108123849

20220108123853

20220108123857

20220108123901

20220108123904

参考文献

[1] 裴明涛, 沈佳峻, 杨敏, 等.复杂光照环境下的车辆检测方法[J].北京理工大学学报, 2016, 36(4):393-398. PEI M T, SHEN J J, YANG M, et al.Vehicle detection method in complex illumination environment[J].Transactions of Beijing Institute of Technology, 2016, 36(4):393-398.(in Chinese)
[2] 周行, 陈淑荣.一种基于多特征融合级联分类器的车辆检测算法[J].现代计算机, 2018(17):38-43. ZHOU X, CHEN S R.A vehicle detection algorithm based on multi-feature fusion and cascade classifier[J].Modern Computer, 2018(17):38-43.(in Chinese)
[3] 黎向锋, 赵伟康, 豆心愿, 等.基于改进AdaBoost+Haar的车辆检测算法[J].测控技术, 2019, 38(2):42-45. LI X F, ZHAO W K, DOU X Y, et al.Vehicle detection algorithm based on improved AdaBoost+Haar[J].Measurement and Control Technology, 2019, 38(2):42-45.(in Chinese)
[4] REDMON J, DIVVALA S, GIRSHICK R, et al.You only look once:unified, real-time object detection[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:779-788.
[5] LIU W, ANGUELOV D, ERHAN D, et al.SSD:single shot MultiBox detector[C]//Proceedings of European Conference on Computer Vision.Berlin, Germany:Springer, 2016:21-37.
[6] ZHANG S, WEN L, BIAN X, et al.Single-shot refinement neural network for object detection[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2018:4203-4212.
[7] GIRSHICK R, DONAHUE J, DARREL T, et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of European Conference on Computer Vision and Pattern Recognition.Berlin, Germany:Springer, 2014:580-587.
[8] GIRSHICK R.Fast R-CNN[C]//Proceedings of European Conference on Computer Vision.Berlin, Germany:Springer, 2015:1440-1448.
[9] REN S, HE K, GIRSHICK R, et al.Faster R-CNN:towards real-time object detection with region proposal networks[C]//Proceedings of NIPSʼ15.Cambridge, USA:MIT Press, 2015:91-99.
[10] 王斌, 顾银娟.基于FPN的航空图像车辆检测技术研究[J].工业工程控制计算机, 2020, 33(7):88-90. WANG B, GU Y Y J.A FPN-based framework for vehicle detection in aerial images[J].Industrial Control Computer, 2020, 33(7):88-90.(in Chinese)
[11] 吴水清, 王宇, 师岩.基于SSD的车辆目标检测[J].计算机与现代化, 2019(5):35-40. WU S Q, WANG Y, SHI Y.Vehicle detection based on SSD[J].Computer and Modernization, 2019(5):35-40.(in Chinese)
[12] 刘肯, 何娇娇, 张永平, 等.改进YOLO的车辆检测算法[J].现代电子技术, 2019, 42(13):47-50. LIU K, HE J J, ZHANG Y P, et al.Improved YOLO vehicle detection algorithm[J].Modern Electronics Technique, 2019, 42(13):47-50.(in Chinese)
[13] SZEGEDY C, LIU W, JIA Y, et al.Going deeper with convolutions[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2015:1-9.
[14] SZGEDY C, VANHOUCKE V, lOFFE S, et al.Rethinking the inception architecture for computer vision[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:2818-2826.
[15] SZGEEDY C, IOFFE S, VANHOUCKE V, et al.Inception-V4, inception-ResNet and the impact of residual connections on learning[C]//Proceedings of IEEE National Conference on Artificial Intelligence.Washington D.C., USA:IEEE Press, 2016:4278-4284.
[16] HU J, SHEN L, SUN G.Squeeze-and-excitation networks[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2018:7132-7141.
[17] CHEN L C, PAPANDREOU G, SCHROFF F, et al.Rethinking astrous convolution for semantic segmentation[EB/OL].[2020-10-12].http://arxiv.org/abs/1706.05587.
[18] LIN T Y, DOLLAR P, GIRSHICK R, et al.Feature pyramid networks for object detection[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:936-944.
[19] LOFFE S, SZEGEDY C.Batch normalization:accelerating deep network training by reducing internal covariate shift[C]//Proceedings of the 32nd International Conference on International Conference on Machine Learning.Washington D.C., USA:IEEE Press, 2015:448-456.
[20] LI Y, DONG H.Multi-block SSD based on small object detection for UAV railway scene surveillance[J].Chinese Journal of Aeronautics, 2020, 33(6):1747-1755.
[21] 郑秋梅, 王璐璐, 王风华.基于改进卷积神经网络的交通场景小目标检测[J].计算机工程, 2020, 46(6):26-33. ZHENG Q M, WANG L L, WANG F H.Small object detection in traffic scene based on improved convolutional neural network[J].Computer Engineering, 2020, 46(6):26-33.(in Chinese)
[22] 储岳中, 黄勇, 张学锋, 等.基于自注意力的SSD图像目标检测算法[J].华中科技大学学报, 2010, 48(9):1-8. CHU Y Z, HUANG Y, ZHANG X F, et al.SSD image target detection algorithm based on self-attention[J].Journal of Huazhong University of Science and Technology, 2010, 48(9):1-8.(in Chinese)
[23] BOCHKOVHKIY A, WANG C, LIAO H.YOLOv4:optimal speed and accuracy of object detection[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2020:198-215.

选择文件类型/文献管理软件名称

选择包含的内容