High-Precision Garbage Detection Algorithm of Lightweight YOLOv5n

doi:10.3778/j.issn.1002-8331.2210-0316

Abstract

Abstract: Aiming at the problems of the existing domestic waste detection models, such as multiple parameters, large amount of calculation, which are not suitable for deployment on mobile devices or embedded devices, and less types of garbage identification, a lightweight and high-precision optimization research is carried out for YOLOv5n target detection algorithm. Firstly, the lightweight networks ShuffleNetv2 and GhostNet are introduced on the YOLOv5n architecture to accomplish the lightweight detection network design. Secondly, the attention mechanism SE is added to enhance the feature extraction ability of the network, and the response-based knowledge distillation algorithm is introduced to improve the accuracy of localization and classification, thereby improving the detection accuracy. Experimental results show that, on the HGI-30 dataset, the optimized YOLOv5n reduces the amount of parameters and computation by 22.3% and 23.3%, and the detection accuracy mAP0.5 and mAP0.5：0.95 are increased by 1.6 percentage points and 2.6 percentage points.

Key words: YOLOv5n, lightweight network, knowledge distillation, domestic waste classification

摘要： 针对现有部署至移动设备或嵌入式设备的生活垃圾检测模型参数量多，计算量大，且识别种类较少等问题，对YOLOv5n目标检测算法进行了轻量化、高精度的优化研究。在YOLOv5n的架构上引入轻量级网络ShuffleNetv2与GhostNet实现了检测网络的轻量化；同时添加注意力机制SE增强特征提取能力，以及引入基于响应的知识蒸馏算法提升定位和分类的准确率，从而提高目标检测精度。实验结果表明，在HGI-30数据集上，优化后的YOLOv5n的参数量和计算量分别减少22.3%和23.3%，检测精度mAP0.5和mAP0.5：0.95分别增加1.6个百分点和2.6个百分点。

关键词: YOLOv5n, 轻量级网络, 知识蒸馏, 生活垃圾分类

TU Chengfeng, YI Anlin, YAO Tao, HE Wenwei. High-Precision Garbage Detection Algorithm of Lightweight YOLOv5n[J]. Computer Engineering and Applications, 2023, 59(10): 187-195.

涂成凤, 易安林, 姚涛, 贺文伟. 轻量化YOLOv5n的高精度垃圾检测算法[J]. 计算机工程与应用, 2023, 59(10): 187-195.

References

[1] HOORNWEG D，BHADA-TATA P.What a waste：a global review of solid waste management[J].Urban Development Series Knowledge Papers，2012，12：27-33.
[2] RABANO S L，CABATUAN M K，SYBINGCO E，et al.Common garbage classification using MobileNet[C]//2018 IEEE 10th International Conference on Humanoid，Nanotechnology，Information Technology，Communication and Control，Environment and Management，Baguio City，Nov 29-Dec 2，2018.Piscataway：IEEE，2018：1-4.
[3] FROST S，TOR B，AGRAWAL R，et al.CompostNet：an image classifier for meal waste[C]//2019 IEEE Global Humanitarian Technology Conference，Seattle，Oct 17-20，2019.Piscataway：IEEE，2019：1-4.
[4] ZHANG Q，ZHANG X，MU X，et al.Recyclable waste image recognition based on deep learning[J].Resources Conservation and Recycling，2021，171：105636.
[5] QIN J M，WANG C，RAN X，et al.A robust framework combined saliency detection and image recognition for garbage classification[J].Waste Management，2022，140：193-203.
[6] CHEN Z H，ZOU H B，WANG Y，et al.Multi-task detection system for garbage sorting base on high-order fusion of convolutional feature hierarchical representation[C]//2018 37th Chinese Control Conference，Wuhan，Jul 25-27，2018.Piscataway：IEEE，2018：5426-5430.
[7] WU H.A YOLOV3 system for garbage detection based on MobileNetV3_Lite as backbone[C]//2021 International Conference on Electronics，Circuits and Information Engineering，Zhengzhou，Jan 22-24，2021.Piscataway：IEEE，2021：254-258.
[8] LI J，CHEN J，SHENG B，et al.Automatic detection and classification system of domestic waste via multimodel cascaded convolutional neural network[J].IEEE Transactions on Industrial Informatics，2021，18（1）：163-173.
[9] KUMSETTY N V，NEKKARE A B，KAMATH S.TrashBox：trash detection and classification using quantum transfer learning[C]//2022 31st Conference of Open Innovations Association，Helsinki，Apr 27-29，2022.Piscataway：IEEE，2022：125-130.
[10] WANG C，ZHOU Y，LI J.Lightweight YOLOv4 target detection algorithm fused with ECA mechanism[J].Processes，2022，10（7）：1285.
[11] LIU C，XIE N，YANG X，et al.A domestic trash detection model based on improved YOLOX[J].Sensors，2022，22（18）：6974.
[12] WU Z，LI H，WANG X，et al.New benchmark for household garbage image recognition[J].Tsinghua Science and Technology，2022，27（5）：793-803.
[13] MA N，ZHANG X，ZHENG H T，et al.ShuffleNet v2：practical guidelines for efficient CNN architecture design[C]//Proceedings of the 15th European Conference on Computer Vision，Munich，Sep 8-14，2018.Cham：Springer，2018：116-131.
[14] HAN K，WANG Y，TIAN Q，et al.GhostNet：more features from cheap operations[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition，Seattle，Jun 13-19，2020.Piscataway：IEEE，2020：1580-1589.
[15] HU J，SHEN L，SUN G.Squeeze-and-excitation networks[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition，Salt Lake City，Jun 18-23，2018.Piscataway：IEEE，2018：7132-7141.
[16] YU G，CHANG Q，LV W，et al.PP-PicoDet：a better real-time object detector on mobile devices[J].arXiv：2111.
00902，2021.
[17] HOU Q，ZHOU D，FENG J.Coordinate attention for efficient mobile network design[C]//Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition，Nashville，Jun 20-25，2021.Piscataway：IEEE，2021：13713-13722.
[18] MEHTA R，OZTURK C.Object detection at 200 frames per second[C]//Proceedings of the 15th European Conference on Computer Vision Workshops，Munich，Sep 8-14，2018.Cham：Springer，2018：659-675.