Traffic Target Detection in Complex Scenes Based on Improved YOLOv5 Algorithm

doi:10.12068/j.issn.1005-3026.2022.08.002

Abstract

Abstract: Real-time target detection in traffic scenarios is the prerequisite of electronic monitoring， automatic driving， and other functions. In view of the low detection efficiency of existing target detection algorithms and the low accuracy of most light target detection algorithms， which are easy to obtain wrong or insufficient target detection， this paper adopts the improved YOLOv5 target detection algorithm for model training， and the pseudo-label strategy for training process optimization. Then， the KITTI traffic target dataset tags are merged into three categories for model training and testing. Through the experimental comparison， the improved YOLOv5 model in this paper achieves 92.5% mAP in all categories， which is 3% higher than the original YOLOv5 training model. Finally， the three categories of the trained models are deployed on the Jetson Nano embedded platform for inference testing， and TensorRT is used to accelerate inference. The average inference time per frame of image is 77ms， which meets the goal of real-time detection.

Key words: deep learning; target detection; YOLOv5 algorithm; pseudo-label training; embedded platform

CLC Number:

TP391

GU De-ying， LUO Yu-lun， LI Wen-chao. Traffic Target Detection in Complex Scenes Based on Improved YOLOv5 Algorithm[J]. Journal of Northeastern University(Natural Science), 2022, 43(8): 1073-1079.

References

[1]Hinton G E，Salakhutdinov R.Reducing the dimensionality of data with neural networks［J］.Science，2006，313(5786):504-507.
[2]Krizhevsky A，Sutskever I，Hinton G E.ImageNet classification with deep convolutional neural networks［C］//Proceedings of the 25th International Conference on Neural Information Processing Systems.New York:ACM，2012:1097-1105.
[3]Girshick R，Donahue J，Darrell T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation［C］//IEEE Conference on Computer Vision and Pattern Recognition.Columbus:IEEE，2014:580-587.
[4]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，2015，39(6):1137-1149.
[5]Tian Y，Gelernter J，Wang X，et al.Lane marking detection via deep convolutional neural network［J］.Neurocomputing，2018，280:46-55.
[6]Redmon J，Divvala S，Girshick R，et al.You only look once:unified，real-time object detection［C］//IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas:IEEE，2016:779-788.
[7]Liu W，Anguelov D，Erhan D，et al.SSD:single shot multi-box detector［M］.Berlin:Springer，2016.
[8]周攀，徐向荣，朱永飞，等.基于改进YOLOv3的道路目标检测［J］.龙岩学院学报，2021，39(2):21-30.(Zhou Pan，Xu Xiang-rong，Zhu Yong-fei，et al.Road target detection based on improved YOLOv3［J］.Journal of Longyan University，2021，39(2):21-30.)
[9]刘力冉，曹杰，杨磊，等.一种改进YOLOv3-Tiny的行车检测算法［J］.计算机与现代化，2020(3):108-114.(Liu Li-ran，Cao Jie，Yang Lei，et al.An improved YOLOv3-Tiny traffic detection algorithm［J］.Computer and Modernization，2020(3):108-114.)
[10]Ma L Z，Chen Y，Zhang J L.Vehicle and pedestrian detection based on improved YOLOv4-Tiny model［J］.Journal of Physics:Conference Series，2021，1920(1):012034.
[11]秦丹峰，尹相辉，龚学余.基于改进YOLOv4-Tiny算法的移动端实时司机违章行为检测［J］.计算机科学与应用，2021，11(5):1291-1300.(Qin Dan-feng，Yin Xiang-hui，Gong Xue-yu.Real-time drivers violation detection on mobile terminal based on improved YOLOv4-Tiny［J］.Computer Science and Applications，2021，11(5):1291-1300.)
[12]Jocher G，Nishimura K，Mineeva T，et al.YOLOv5［EB/OL］.(2020-06-26)［2021-06-02］.https://github.com/ultralytics/yolov5.
[13]He K，Zhang X，Ren S，et al.Deep residual learning for image recognition［EB/OL］.(2015-12-10)［2021-06-02］.https://arxiv.org/abs/1512.03385.
[14]Trinh H C，Lee D H，Kwon Y K.PANET:a GPU based tool for fast parallel analysis of robustness dynamics and feed-forward/ feedback loop structures in large-scale biological networks［J］.PLoS ONE，2014，9(7):e103010.
[15]Rezztofighi H，Tsoi N，Gwak J Y，et al.Generalized intersection over union:a metric and a loss for bounding box regression［EB/OL］.(2019-04-15)［2021-06-02］.https://arxiv.org/pdf/1902.09630.
[16]Jocher G，Nishimura K，Mineeva T，et al.v5.0-YOLOv5-P6.1280 models，AWS，supervise.ly and YouTube integrations［EB/OL］.(2021-04-12)［2021-06-02］.http://github.com/ultralytics/yolov5releases/tag/v5.0.
[17]Geiger A，Lenz P，Urtasun R，et al.Are we ready for autonomous driving? The KITTI vision benchmark suite［C］// IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas:IEEE，2012:3354-3361.