| 4,663 | 17 | 25 |
| 下载次数 | 被引频次 | 阅读次数 |
为了解决车辆目标检测中准确率低的问题，提出了一种基于改进YOLOv5算法的车辆目标检测.改进后的YOLOv5算法主要是在原来的基础上通过K-means聚类的方法对数据集中的目标边框进行重新聚类、并将CIoU损失函数和DIoU＿nms应用于YOLOv5算法来提高目标识别效果.改进后的YOLOv5算法，目标检测mAP达到了85.8%,比改进前的YOLOv5算法提升了1.3%.
Abstract：In the automatic driving system, vehicle target detection is a key content and basic task. In order to ensure road safety, it is necessary to accurately detect all targets on the road. In order to solve the problem of low accuracy in vehicle target detection, the article proposes a vehicle target detection algorithm based on improved YOLOv5. The improved YOLOv5 algorithm mainly re-clusters the target borders in the data set through the K-means clustering method on the original basis, and applies the CIoU loss function and DIoU＿nms to the YOLOv5 network to improve the target recognition effect. With the improved YOLOv5 algorithm, the target detection mAP reached 85.8%,which is 1.3% higher than the previous YOLOv5 algorithm.
[1] GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich feature hierarchies for accurate object detection and semanticse gmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Patternr-ecognition,2014:580-587.
[2] GIRSHICKR.FastR-CNN[C]//2015 IEEE International Conference on Computer Vision(ICCV).New York:IEEE,2015:1440-1448.
[3] RENSQ,HEKM,GIRSHICKR,et al.Faster R-CNN:Towards real-time object detection with region proposal networks[C]//IEEE Transaction Pattern Analysisand Machine Intelligence.New York:IEEE,2019:1137-1149.
[4] LIU W,ANGUELOV D,ERHAN D,et al.Ssd:Single shot multibox detector[C]//European Conference on Computer Vision.Springer,Cham,2016:21-37.
[5] 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,Las Vegas,779-788.
[6] REDMON J,FARHADI A.YOLO9000:Better,faster,stronger[C]//Proceedings of the IEEE conference on computer vision and pattern recognition.2017:7263-7271.
[7] REDMON J,FARHADI A.Yolov3:An incremental improvement[J].arXiv preprint arXiv:1804.02767,2018.
[8] 刘云霄，王敬东，黄雨秋，等 .MTCNN 的改进及其在道路车辆检测中的应用[J].光电子技术，2019,39(3):196-204,224.
[9] 王聪 .基于深度学习的无人机单目标识别与跟踪算法研究[D].厦门：华侨大学，2019:14-16.
[10] REDMON J,FARHADI A.YOLO9000:Better,faster,stronger[C]// IEEE Conference on Computer Vision & Pattern Recognition.IEEE,2017:6517-6525.
[11] 张麒麟，林清平，肖蕾.改进YOLOv5的航拍图像识别算法[J].长江信息通信，2021,34(3):73-76.
[12] TAN Shilei,LU Gonglin,JIANG Ziqiang,et al.Improved YOLOv5 network model and application in safety helmet detection[C]//Proceedings of the 2021 IEEE International Conference on Intelligence and Safety for Robotics.Nagoya,Japan.
[13] BOCHKOVSKIY A,WANG C Y,LIAO H.YOLOv4:Optimal speed and accuracy of object detection[J].arXiv preprint arXiv:2004.10934,2020.
[14] 张洋.改进的YOLOv3-tiny在城市交叉路口车辆检测中的应用[D].重庆：重庆师范大学，2020.DOI:10.27672/d.cnki.gcsfc.2020.000818.
[15] DAI J,LI Y,HE K,et al.R-fcn:Object detection via region-based fully convolutional networks[C]//Advances in neural information processing systems.2016:379-387.
基本信息:
DOI：
中图分类号:TP391.41;TP183;U463.6
引用信息:
[1]刘超阳,曲金帅,范菁等.基于改进YOLOv5算法的车辆目标检测[J].云南民族大学学报(自然科学版),2022,31(06):749-754.
基金信息: