基于改进YOLOX的轻量型垃圾分类检测方法

doi:10.16088/j.issn.1001-6600.2022100804

摘要/Abstract

摘要： 生活垃圾分类是保护生态环境、促进绿色和谐发展的有效措施。针对移动端设备计算资源和内存有限,重量级模型难以嵌入等问题,本文提出一种基于改进YOLOX-tiny轻量型的垃圾分类检测方法。首先,使用EIoU替换原来的IoU损失函数,能加速收敛,提升检测精度;其次,在颈部网络引入注意力机制CBAM,对不同通道的权重重新分配,获取更多浅层的细粒度特征和深层的语义信息;最后,使用GhostBottleneck模块替换特征提取网络中的CSP模块,保留更多边缘信息,同时降低参数量,使模型轻量化。在华为云垃圾数据集上的实验结果表明,改进的算法与YOLOX-tiny相比,参数量降低至原来的87.97%,精度提升了0.3个百分点,在TrashNet数据集上的实验效果提升了0.36个百分点,从而证明了本文算法的有效性,该算法有利于嵌入移动端设备使用,具有一定的实用价值。

关键词: 垃圾分类, YOLOX, 轻量型网络, EIoU, CBAM, GhostBottleneck

Abstract: Household garbage classification is an effective measure to protect the ecological environment and promote green and harmonious development. Aiming at the problems such as limited computing resources and memory, and difficulty in embedding heavyweight models into mobile devices, a lightweight garbage classification detection method based on improved YOLOX-tiny is proposed in this paper. Firstly, the original IoU loss function is replaced by EIoU, which can accelerate the convergence and improve the detection accuracy. Secondly, the attention mechanism CBAM is introduced into the neck network to redistribute the weight of different channels to obtain more shallow fine-grained features and deep semantic information. Finally, the GhostBottleneck module is used to replace the CSP module in the feature picking network, which tends to retain more edge information, reduce the number of parameters, and lighten the model. Experimental results on Huawei cloud garbage dataset show that compared with YOLOX-tiny, the number of parameters of the improved algorithm is reduced to 87.97% of the original, the accuracy is increased by 0.3%, and the experimental effect on TrashNet dataset is increased by 0.36%, which proves the effectiveness of the proposed algorithm. The algorithm is conducive to the use of embedded mobile devices and has certain practical value.

Key words: garbage classification, YOLOX, lightweight network, EIoU, CBAM, GhostBottleneck

中图分类号: TP391.41

李洋, 苟刚. 基于改进YOLOX的轻量型垃圾分类检测方法[J]. 广西师范大学学报（自然科学版）, 2023, 41(3): 80-90.

LI Yang, GOU Gang. Lightweight Garbage Detection Method Based on Improved YOLOX[J]. Journal of Guangxi Normal University(Natural Science Edition), 2023, 41(3): 80-90.

参考文献

[1] 国家统计局. 中国统计年鉴2021[M]. 北京: 中国统计出版社, 2021.
[2] 张涛, 白冬锐, 孙煜璨, 等. 全过程管理视角的上海市垃圾分类回顾与展望[J]. 环境工程, 2022, 40(3): 173-180, 146. DOI: 10.13205/j.hjgc.202203026.
[3] 王洁, 顾卫华, 陈泽辉, 等. 生活垃圾分类实践效果、问题与对策分析: 以湖州市织里镇为例[J]. 环境工程, 2022, 40(3): 188-193. DOI: 10.13205/j.hjgc.202203028.
[4] 贵阳市城镇生活垃圾分类管理条例[N]. 贵阳日报, 2022-08-16(4).
[5] 李永杰, 周桂红, 刘博. 基于YOLOv3模型的人脸检测与头部姿态估计融合算法[J]. 广西师范大学学报(自然科学版), 2022, 40(3): 95-103. DOI: 10.16088/j.issn.1001-6600.2021070911.
[6] 刘英璇, 伍锡如, 雪刚刚. 基于深度学习的道路交通标志多目标实时检测[J]. 广西师范大学学报(自然科学版), 2020, 38(2): 96-106. DOI: 10.16088/j.issn.1001-6600.2020.02.011.
[7] 吕方方, 陈光喜, 刘家畅,等. 基于卷积神经网络的小目标检测改进算法[J]. 桂林电子科技大学学报, 2021, 41(5): 368-374. DOI: 10.16725/j.cnki.cn45-1351/tn.2021.05.005.
[8] GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2014: 580-587. DOI: 10.1109/CVPR.2014.81.
[9] GIRSHICK R. Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision (ICCV). Los Alamitos, CA: IEEE Computer Society, 2015: 1440-1448. DOI: 10.1109/ICCV.2015.169.
[10] REN S Q, HE K M, 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. DOI: 10.1109/TPAMI.2016.2577031.
[11] LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Computer Vision-ECCV2016: LNCS Volume 9905. Cham: Springer, 2016: 21-37. DOI: 10.1007/978-3-319-46448-0_2.
[12] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2016: 779-788. DOI: 10.1109/CVPR.2016.91.
[13] REDMON J, FARHADI A. YOLO9000: better, faster, stronger[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2017: 6517-6525. DOI: 10.1109/CVPR.2017.690.
[14] REDMON J, FARHADI A. YOLOv3: an incremental improvement[EB/OL]. (2018-04-08)[2022-10-08]. https://arxiv.org/abs/1804.02767. DOI: 10.48550/arXiv.1804.02767.
[15] GE Z, LIU S T, WANG F, et al. YOLOX: exceeding YOLO series in 2021[EB/OL]. (2021-08-06)[2022-10-08]. https://arxiv.org/abs/2107.08430. DOI: 10.48550/arXiv.2107.08430.
[16] 陈智超, 焦海宁, 杨杰, 等. 基于改进MobileNet v2的垃圾图像分类算法[J]. 浙江大学学报(工学版), 2021, 55(8): 1490-1499. DOI: 10.3785/j.issn.1008-973X.2021.08.010.
[17] 高明, 陈玉涵, 张泽慧, 等. 基于新型空间注意力机制和迁移学习的垃圾图像分类算法[J]. 系统工程理论与实践, 2021, 41(2): 498-512. DOI: 10.12011/SETP2020-1645.
[18] 袁建野, 南新元, 蔡鑫, 等. 基于轻量级残差网路的垃圾图片分类方法[J]. 环境工程, 2021, 39(2): 110-115. DOI: 10.13205/j.hjgc.202102017.
[19] 罗安能, 万海斌, 司志巍, 等. 基于改进YOLOv5s的可回收垃圾的检测算法[J/OL]. 激光与光电子学进展: 1-15[2022-10-08]. http://kns.cnki.net/kcms/detail/31.1690.tn.20220713.1957.657.html.
[20] 吕东, 王萍, 王宇航, 等. 固体金属垃圾分类中基于深度学习方法的研究[J]. 广西科技大学学报, 2021, 32(4): 104-110, 126. DOI: 10.16375/j.cnki.cn45-1395/t.2021.04.016.
[21] YU J H, JIANG Y N, WANG Z Y, et al. Unitbox: an advanced object detection network[C]//Proceedings of the 24th ACM International Conference on Multimedia. New York, NY: Association for Computing Machinery, 2016: 516-520. DOI: 10.1145/2964284.2967274.
[22] REZATOFIGHI H, TSOI N, GWAK J Y, et al. Generalized intersection over union: a metric and a loss for bounding box regression[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2019: 658-666. DOI: 10.1109/CVPR.2019.00075.
[23] ZHENG Z H, WANG P, LIU W, et al. Distance-IoU loss: faster and better learning for bounding box regression[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(7): 12993-13000. DOI: 10.1609/aaai.v34i07.6999.
[24] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2018: 7132-7141. DOI: 10.1109/CVPR.2018.00745.
[25] WOO S H, PARK J C, LEE J Y, et al. CBAM: convolutional block attention module[C]//Computer Vision-ECCV 2018: LNCS Volume 11211. Cham: Springer Nature Switzerland AG, 2018: 3-19. DOI: 10.1007/978-3-030-01234-2_1.
[26] HAN K, WANG Y H, TIAN Q, et al. GhostNet: more features from cheap operations[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2020: 1577-1586. DOI: 10.1109/CVPR42600.2020.00165.
[27] HOWARD A, SANDLER M, CHEN B, et al. Searching for MobileNetV3[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV). Los Alamitos, CA: IEEE Computer Society, 2019: 1314-1324. DOI: 10.1109/ICCV.2019.00140.
[28] 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. DOI: 10.1016/j.neucom.2022.07.042.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed