基于改进YOLOv3的高速公路隧道内停车检测方法

doi:10.3778/j.issn.1002-8331.2107-0033

摘要/Abstract

摘要：

为了更准确地检测高速公路隧道内停车行为，提出一种基于改进YOLOv3车辆检测模型的高速公路隧道内停车检测方法。通过筛选VOC数据集以及实际高速公路隧道内的车辆图片制作专门用于高速公路隧道内车辆检测的数据集，选取YOLOv3目标检测模型作为车辆检测的基础网络结构，并对其进行加深网络结构的改进使其能够准确检测隧道内的车辆。将Deep SORT跟踪算法应用于改进的停车检测模型中，对车辆进行跟踪从而计算行驶速度，并创新性地设置双重速度阈值来判别车辆的停车行为。实验结果表明，经过改进的YOLOv3模型相比于原模型，在VOC-vehicle数据集和Tunnel-vehicle数据集上的mAP都有所提升，最终获得了mAP为98.19%的高速公路隧道车辆检测模型。将基于改进YOLOv3的高速公路隧道内停车检测方法在高速公路隧道视频上进行测试，可以有效地在高速公路隧道中完成停车检测的任务。

关键词: 深度学习, YOLOv3, 高速公路隧道, 车辆跟踪, 停车检测

Abstract:

In order to detect the vehicle stopping behaviors more accurately in highway tunnel, a detection method based on the improved YOLOv3 is proposed. Firstly, a highway tunnel data set is produced and the YOLOv3 object detection model is deepened to improve the network structure and accurately detect vehicles in the tunnel. Then the Deep SORT algorithm is used to track the vehicle, and the vehicle speed is calculated. The speed threshold is used to determine whether a stopping behavior has occurred. The experimental results show that, compared with the original model, the optimized YOLOv3 model in this paper improves the mAP on VOC-vehicle data set and Tunnel-vehicle data set. Finally, a vehicle detection model with 98.19% mAP is obtained. In addition, based on the improved YOLOv3, the detection method of highway tunnel vehicle stopping is tested on the highway tunnel video and the results show that the detection of the vehicle stopping behaviors can be completed effectively.

Key words: deep learning, YOLOv3, highway tunnel, vehicle tracking, vehicle stopping detection

丁冰，杨祖莨，丁洁，刘晋峰，闫国亮. 基于改进YOLOv3的高速公路隧道内停车检测方法[J]. 计算机工程与应用, 2021, 57(23): 234-239.

DING Bing, YANG Zuliang, DING Jie, LIU Jinfeng, YAN Guoliang. Detection Method of Highway Tunnel Vehicle Stopping Based on Improved YOLOv3[J]. Computer Engineering and Applications, 2021, 57(23): 234-239.

参考文献

[1] 张亮兵.高速公路隧道监控系统的组成与作用[J].科技创新与应用，2017（16）：233.
ZHANG L B.The composition and function of highway tunnel monitoring system[J].Technological Innovation and Application，2017（16）：233.
[2] 晋济高速岩后隧道“3·1”事故处理结果公布[J].专用汽车，2014（7）：76.
The results of the “3·1” accident in the Yanhou Tunnel of Jinji Expressway are announced[J].Special Vehicle，2014（7）：76.
[3] GULER S，SILVERSTEIN J A，PUSHEE I H.Stationary objects in multiple object tracking[C]//2007 IEEE Conference on Advanced Video and Signal Based Surveillance，2007：248-253.
[4] PORIKLI F.Detection of temporarily static regions by processing video at different frame rates[C]//2007 IEEE Conference on Advanced Video and Signal Based Surveillance，2007：236-241.
[5] 赵敏，张波，唐毅，等.露天场景下高速公路违章停车视频检测算法[J].计算机测量与控制，2014，22（2）：489-492.
ZHAO M，ZHANG B，TANG Y，et al.Freeway illegal parking video detection algorithm in open air scene[J].Computer Measurement and Control，2014，22（2）：489-492.
[6] 杨祖莨，丁洁，刘晋峰.一种新的结合卷积神经网络的隧道内停车检测方法[J].重庆大学学报，2021，44（6）：49-59.
YANG Z L，DING J，LIU J F.A new tunnel parking detection method combined with convolutional neural network[J].Journal of Chongqing University，2021，44（6）：49-59.
[7] 于青青，张可，唐恒亮，等.基于视频跟踪轨迹的全过程路侧停车行为检测与识别技术[J].计算机与现代化，2017（9）：67-73.
YU Q Q，ZHANG K，TANG H L，et al.A roadside parking detection and recognition method based on video motion trail[J].Computer and Modernization，2017（9）：67-73.
[8] 姜明新，王洪玉，沐方顺.基于视频监控的违章停车检测算法[J].计算机工程，2012，38（19）：151-153.
JIANG M X，WANG H Y，MU F S.Illegal parking detection algorithm based on video surveillance[J].Computer Engineering，2012，38（19）：151-153.
[9] 石时需，郑启伦，黄翰，等.基于粒子滤波算法的高速公路车辆停车检测[J].计算机工程与应用，2008，44（34）：239-242.
SHI S X，ZHENG Q L，HUANG H，et al.Vehicle breaking detection on express way based on particle filter algorithm[J].Computer Engineering and Applications，2008，44（34）：239-242.
[10] 杜金航，何宁.基于改进的YOLOv3道路车辆实时检测[J].计算机工程与应用，2020，56（11）：26-32.
DU J H，HE N.Real-time road vehicles detection based on improved YOLOv3[J].Computer Engineering and Applications，2020，56（11）：26-32.
[11] REDMON J，FARHADI A.Yolov3：an incremental improvement[J].arXiv：1804.02767，2018.
[12] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：779-788.
[13] REDMON J，FARHADI A.YOLO9000：better，faster，stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：7263-7271.
[14] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[C]//Advances in Neural Information Processing Systems，2015：91-99.
[15] EVERINGHAM M，ESLAMI S M A，VAN GOOL L，et al.The pascal visual object classes challenge：a retrospective[J].International Journal of Computer Vision，2015，111（1）：98-136.
[16] REDMON J.Darknet：open source neural networks in C[Z].2013.
[17] 张富凯，杨峰，李策.基于改进YOLOv3的快速车辆检测方法[J].计算机工程与应用，2019，55（2）：12-20.
ZHANG F K，YANG F，LI C.Fast vehicle detection method based on improved YOLOv3[J].Computer Engineering and Applications，2019，55（2）：12-20.
[18] VEERAMANI B，RAYMOND J W，CHANDA P.DeepSort：deep convolutional networks for sorting haploid maize seeds[J].BMC Bioinformatics，2018，19（9）：289.
[19] CHEN X，FANG H，LIN T Y，et al.Microsoft COCO captions：data collection and evaluation server[J].arXiv：1504.00325，2015.
[20] HE K，SUN J，TANG X.Single image haze removal using dark channel prior[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2011，33（12）：2341-2353.