基于YOLOv5s室内目标检测轻量化改进算法研究

doi:10.3778/j.issn.1002-8331.2305-0109

摘要/Abstract

摘要： 针对现有室内目标检测算法，存在结构复杂，计算量以及模型参数量过大等问题，难以部署到计算能力有限的室内机器人平台，实现高效的目标检测。为解决这一问题，提出了一种改进的YOLOV5s轻量化检测算法。该方法采用ShuffleNetv2作为主干特征提取网络，并且在改进的主干网络基础上采用CA注意力机制，同时在颈部网络中采用GSConv和VOV-GSCSP模块。最后引入边框回归损失函数EIOU加快网络收敛。研究结果表明，改进后的目标检测算法，模型计算量减少了68.75%，模型参数量减少了62.2%，权重文件减少了59.7%，平均精确率mAP均值为0.653，改进后的目标检测模型能够在保证轻量化的同时保证检测精度。

关键词: YOLOv5s, 轻量化, ShuffleNetv2网络, CA注意力机制, GSConv模块, VOV-GSCSP模块, EIOU损失函数

Abstract: The existing indoor target detection algorithms have many problems, such as complex structure, large amount of calculation and large number of model parameters, which are difficult to be deployed to the indoor robot platform with limited computing capacity to achieve efficient target detection. To solve this problem, an improved YOLOV5s detection algorithm is proposed. In this method, ShuffleNetv2 is introduced as the backbone feature extraction network, and CA attention mechanism is adopted on the basis of the improved backbone network, and GSConv and VOV-GSCSP modules are adopted in the neck network. Finally, the bounding regression loss function EIOU is introduced to accelerate the network convergence. The results show that the improved target detection algorithm reduces the model computation by 68.75%, the number of model parameters by 62.2%, the weight file by 59.7%, and the average accuracy mAP is 0.653. The improved target detection model can ensure the detection accuracy while ensuring the lightweight.

Key words: YOLOv5s, light weight, ShuffleNetv2, CA attention mechanism, GSConv module, VOV-GSCSP module, EIOU loss function

牛鑫宇, 毛鹏军, 段云涛, 娄晓恒. 基于YOLOv5s室内目标检测轻量化改进算法研究[J]. 计算机工程与应用, 2024, 60(3): 109-118.

NIU Xinyu, MAO Pengjun, DUAN Yuntao, LOU Xiaoheng. Research on Lightweight Improved Algorithm for Indoor Target Detection Based on YOLOv5s[J]. Computer Engineering and Applications, 2024, 60(3): 109-118.

参考文献

[1] 徐凌伟, 权天祺. 基于BP神经网络的移动安全性能预测[J]. 聊城大学学报(自然科学版), 2020, 33(3): 34-40.
XU L W, QUAN T Q. Mobile secrecy performance prediction based on BP neural network[J]. Journal of Liaocheng University (Natural Sciencen), 2020, 33(3): 34-40.
[2] VIOLA P, JONES M. Rapid object detection using a boosted cascade of simple features[C]//Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2001), 2001.
[3] DALAL N, TRIGGS B. Histograms of oriented gradients for human detection[C]//2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), 2005: 886-893.
[4] FELZENSZWALB P F, GIRSHICK R B, MCALLESTER D, et al. Object detection with discriminatively trained part-based models[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 32(9): 1627-1645.
[5] 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, 2014: 580-587.
[6] GIRSHICK R. Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision, 2015: 1440-1448.
[7] 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.
[8] 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.
[9] REDMON J, FARHADI A. YOLO9000: better, faster, stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 7263-7271.
[10] REDMON J, FARHADI A. YOLOv3: an incremental improvement[J]. arXiv:1804.02767, 2018.
[11] BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: optimal speed and accuracy of object detection[J]. arXiv:2004.10934, 2020.
[12] LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Proceedings 14th European Conference on Computer Vision (ECCV 2016), Amsterdam, The Netherlands, October 11-14, 2016: 21-37.
[13] 叶鲁斌, 刘皓月, 张纪元, 等. 家居环境中的智能扫地机器人设计[J]. 计算机辅助设计与图形学学报, 2023, 35(2): 262-272.
YE L B, LIU H Y, ZHANG J Y, et al. Design and implementation of AI algorithms in home-cleaning robot [J]. Journal of Computer-Aided Design and Computer Graphics, 2023, 35(2): 262-272.
[14] LV Y, LIU J, CHI W, et al. An inverted residual based lightweight network for object detection in sweeping robots[J]. Applied Intelligence, 2022, 52(11): 12206-12221.
[15] ZHUANG Y, LIN X, HU H, et al. Using scale coordination and semantic information for robust 3-D object recognition by a service robot[J]. IEEE Sensors Journal, 2014, 15(1): 37-47.
[16] CHEN Z, YANG J, CHEN L, et al. Garbage classification system based on improved ShuffleNet v2[J]. Resources, Conservation and Recycling, 2022, 178: 106090.
[17] YING B, XU Y, ZHANG S, et al. Weed detection in images of carrot fields based on improved YOLOv4[J]. Traitement du Signal, 2021, 38(2): 341-348.
[18] SU F, ZHAO Y, WANG G, et al. Tomato maturity classification based on SE-YOLOv3-MobileNetV1 network under nature greenhouse environment[J]. Agronomy, 2022, 12(7): 1638.
[19] GUO K, HE C, YANG M, et al. A pavement distresses identification method optimized for YOLOv5s[J]. Scientific Reports, 2022, 12(1): 3542.
[20] 何雨, 田军委, 张震, 等. YOLOv5目标检测的轻量化研究[J]. 计算机工程与应用, 2023, 59(1): 92-99.
HE Y, TIAN J W, ZHANG Z, et al. Lightweight research of YOLOv5 target detection [J]. Computer Engineering and Applications, 2023, 59(1): 92-99.
[21] MA N, ZHANG X, ZHENG H T, et al. Shufflenet v2: practical guidelines for efficient CNN architecture design[C]//Proceedings of the European Conference on Computer Vision (ECCV), 2018: 116-131.
[22] LI H, LI J, WEI H, et al. Slim-neck by GSConv: a better design paradigm of detector architectures for autonomous vehicles[J]. arXiv:2206.02424, 2022.
[23] ZHANG Y F, REN W, ZHANG Z, et al. Focal and efficient IOU loss for accurate bounding box regression[J]. Neurocomputing, 2022, 506: 146-157.
[24] 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, 2017: 2980-2988.
[25] 张学立, 贾新春, 王美刚, 等. 安全帽与反光衣的轻量化检测: 改进YOLOv5s的算法[J]. 计算机工程与应用, 2024, 60(1): 104-109.
ZHANG X L, JIA X C, WANG M G, et al. Lightweight detection of helmets and reflective clothings: improved YOLOv5s algorithm[J]. Computer Engineering and Applications, 2024, 60(1): 104-109.
[26] WANG X, WU Z, JIA M, et al. Lightweight SM-YOLOv5 tomato fruit detection algorithm for plant factory[J]. Sensors, 2023, 23(6): 3336.