改进YOLOX的自动驾驶场景目标检测算法

doi:10.3778/j.issn.1002-8331.2310-0197

摘要/Abstract

摘要： 针对原始YOLOX在自动驾驶场景中易出现小目标的误检及漏检问题，提出一种改进YOLOX网络模型的检测方法。增加一个检测头，显著提高了小目标的检测精度。用更快的空间金字塔池化（spatial pyramid pooling-fast，SPPF）替代传统的空间金字塔池化（spatial pyramid pooling，SPP）结构，并引入了高效通道注意力机制（efficient channel attention，ECA），以增强模型对复杂背景和小目标的识别能力，有效降低了小目标的漏检率。在网络的neck部分，采用深度可分离的卷积方法，进一步提高了计算效率。基于KITTI数据集，设计了一种新的数据增强技术，提高模型的训练稳定性。实验结果表明，优化后YOLOX算法的小目标平均精度APS和平均召回率AR分别提升了0.20和0.097，在小目标检测上取得了显著进步，并大大降低了漏检率。

关键词: 自动驾驶, YOLOX, 目标检测, 注意力机制

Abstract: Aiming at the problem that the original YOLOX is prone to false detection and missed detection of small objects in autonomous driving scenarios, a detection method based on improved YOLOX network model is proposed. Firstly, a detection head is added to improve the detection accuracy of small objects significantly. Secondly, the traditional spatial pyramid pooling (SPP) structure is replaced by spatial pyramid pooling fast (SPPF). Moreover, the efficient channel attention (ECA) mechanismis introduced to enrich feature representation without increasing network complexity and reduce the missed detection rate of small targets effectively. In the neck portion of the network, depthwise separable convolutions are employed to further boost computational efficiency. Finally, based on the KITTI dataset, a novel data augmentation technique is designed to enhance the training stability of the model. Experimental results indicate that the APS and AR of the optimized YOLOX algorithm are increased by 0.20 and 0.097 respectively, which makes significant progress in small object detection and greatly reduces the missed detection rate.

Key words: autonomous driving, YOLOX, object detection, attention mechanism

陈商盈, 倪受东, 童林. 改进YOLOX的自动驾驶场景目标检测算法[J]. 计算机工程与应用, 2024, 60(12): 225-233.

CHEN Shangying, NI Shoudong, TONG Lin. Improved YOLOX Algorithm for Object Detection in Autonomous Driving Scenarios[J]. Computer Engineering and Applications, 2024, 60(12): 225-233.

参考文献

[1] ASKELAND J, LEVINSON J, BECKER J, et al. Towards fully autonomous driving systems and algorithms[C]//IEEE Intelligent Vehicles Symposium (IV), Baden-Baden, Jun 5-9, 2011. New York: IEEE, 2011: 163-168.
[2] JANAI J, GüNEY F, BEHL A, et al. Computer vision for autonomous vehicles: problems, datasets and state of the art[J]. Foundations and Trends? in Computer Graphics and Vision, 2020, 12(1/3): 1-308.
[3] MAHAUR B, MISHRA K K. Small-object detection based on YOLOv5 in autonomous driving systems[J]. Pattern Recognition Letters, 2023, 168: 115-122.
[4] ZHOU Y, WEN S J, WANG D L, et al. MobileYOLO: real-time object detection algorithm in autonomous driving scenarios[J]. Sensors, 2022, 22(9): 3349.
[5] 余以春, 李明旭. 改进YOLOv5s的自动驾驶汽车目标检测[J]. 计算机系统应用, 2023, 32(9): 97-105.
YU Y C, LI M X. Improved YOLOv5s for autonomous vehicle target detection[J]. Computer Systems and Applications, 2023, 32(9): 97-105.
[6] 任钰. 基于Faster R-CNN的小目标检测研究与应用[D]. 安庆: 安庆师范大学, 2022.
REN Y. Research and application of small object detection based on Faster R-CNN[D]. Anqing: Anqing Normal University, 2022.
[7] WANG H, XU Y S, WANG Z N, et al. Centernet-auto: a multi?object visual detection algorithm for autonomous driving scenes based on improved centernet[J]. IEEE Transactions on Emerging Topics in Computational Intelligence, 2023, 7(3): 742-752.
[8] 谢若冰, 李茂军, 李宜伟, 等. 改进YOLOX-s的密集垃圾检测方法[J]. 计算机工程与应用, 2024, 60(5): 250-258.
XIE R B, LI M J, LI Y W, et al. Improving YOLOX-s dense garbage detection method[J]. Computer Engineering and Applications, 2024, 60(5): 250-258.
[9] 王鹏, 王玉林, 焦博文, 等. 基于YOLOv5的道路目标检测算法研究[J]. 计算机工程与应用, 2023, 59(1): 117-125.
WANG P, WANG Y L, JIAO B W, et al. Research on road target detection algorithm based on YOLOv5[J]. Computer Engineering and Applications, 2023, 59(1): 117-125.
[10] CAO Y N, LI C, PENG Y K, et al. MCS-YOLO: a multiscale object detection method for autonomous driving road environment recognition[J]. IEEE Access, 2023, 11: 22342-22354.
[11] WOO S, PARK J, LEE J Y, et al. Cbam: convolutional block attention module[C]//Proceedings of the European Conference on Computer Vision (ECCV), Munich, Sep 8-14, 2018. Berlin: Springer, 2018: 3-19.
[12] WU S Y, YAN Y B, WANG W Q. CF-YOLOx: an autonomous driving detection model for multi-scale object detection[J]. Sensors, 2023, 23(8): 3794.
[13] LI G F, FAN W Q, XIE H, et al. Detection of road objects based on camera sensors for autonomous driving in various traffic situations[J]. IEEE Sensors Journal, 2022, 22(24): 24253-24263.
[14] WANG Q, WU B, ZHU P, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, Jun14-19, 2020. Piscataway, NJ: IEEE, 2020: 11534-11542.
[15] 陈垦, 欧鸥, 杨长志, 等. 基于改进YOLOX的落石检测方法[J]. 计算机测量与控制, 2023, 31(11): 53-59.
CHEN K, OU O, YANG C Z, et al. Rockfall detection method based on improved YOLOX[J]. Computer Measurement & Control, 2023, 31(11): 53-59.
[16] 刘昱杉, 刘卫康, 刘庆华, 等. 基于YOLOX结合DeepSort的船载车辆行人检测算法[J]. 计算机与现代化, 2023(8): 60-67.
LIU Y S, LIU W K, LIU Q H, et al. Pedestrian detection algorithm for ship?borne vehicles based on YOLOX combined with DeepSort[J]. Computer and Modernization, 2023(8): 60-67.
[17] 韩锟, 彭晶莹. 基于改进YOLOx与多级数据关联的行人多目标跟踪算法研究[J]. 铁道科学与工程学报, 2024, 21(1): 94-105.
HAN K, PENG J Y. Pedestrian multi-object tracking algorithm based on improved YOLOx and multi-level data association[J]. Journal of Railway Science and Engineering, 2024, 21(1): 94-105.
[18] 娄树理, 王岩, 郭建勤, 等. 改进YOLOX-S的红外舰船目标检测算法[J]. 应用光学, 2023, 44(5): 1054-1060.
LOU S L, WANG Y, GUO J Q, et al. Infrared ship target detection algorithm based on improved YOLOX-S[J]. Journal?of?Applied?Optics, 2023, 44(5): 1054-1060.
[19] GE Z, LIU S, WANG F, et al. Yolox: exceeding yolo series in 2021[J]. arXiv:2107.08430, 2021.
[20] HE K, ZHANG X, REN S, 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.
[21] CHOLLET F. Xception: deep learning with depthwise separable convolutions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Hawaii, Jul 21-26, 2017. Piscataway, NJ: IEEE, 2017: 1251-1258.
[22] GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Columbus, Jun 23-28, 2014. Piscataway, NJ: IEEE, 2014: 580-587.
[23] 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, 2017, 39(6): 1137-1149.
[24] LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Proceedings of the European Conference on Computer Vision (ECCV), Amsterdam, Oct 10-16. Berlin: Springer, 2016: 21-37.
[25] 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 (CVPR), Las Vegas, Jun 26-Jul 1, 2016. Piscataway, NJ: IEEE, 2016: 779-788.
[26] BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: optimal speed and accuracy of object detection[J]. arXiv:2004.10934, 2020.
[27] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, Jun 26-Jul 1, 2016. Piscataway, NJ: IEEE, 2016: 770-778.
[28] LIN T Y, DOLLAR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR), Hawaii, Jul 21-26, 2017. Piscataway, NJ: IEEE, 2017: 2117-2125.
[29] LIU S, QI L, QIN H, et al. Path aggregation network for instance segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, Jun 18-22, 2018. Piscataway, NJ: IEEE, 2018: 8759-8768.
[30] HU J, SHEN L, ALBANIE S, et al. Squeeze-and-excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, Jun 18-22, 2018. Piscataway, NJ: IEEE, 2018: 7132-7141.
[31] XU L Z, YAN W, JI J S. The research of a novel wog-yolo algorithm for autonomous driving object detection[J]. Scientific Reports, 2023, 13(1): 3699.
[32] GEIGER A, LENZ P, STILLER C, et al. Vision meets robotics: the kitti dataset[J]. The International Journal of Robotics Research, 2013, 32(11): 1231-1237.
[33] ZHANG H, CISSE M, DAUPHIN Y N, et al. Mixup: beyond empirical risk minimization[J]. arXiv:1710.09412, 2017.
[34] 胡淼, 姜麟, 陶友凤, 等. 改进YOLOv7的自动驾驶目标检测算法[J]. 计算机工程与应用, 2024, 60 (11): 165-172.
HU M, JIANG L, TAO Y F, et al. Improved YOLOv7 automatic driving object detection algorithm[J]. Computer Engineering and Applications, 2024, 60 (11): 165-172.
[35] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Oct 22-29, 2017. Piscataway, NJ: IEEE, 2017: 2980-2988.
[36] WANG C Y, YEH I H, LIAO H Y M. You only learn one representation: unified network for multiple tasks[J]. arXiv: 2105.04206, 2021.
[37] REDMON J, FARHADI A. Yolov3 an incremental improvement[J]. arXiv:1804.02767, 2018.