基于YOLO v5模型的非住宅区自动垃圾分类箱设计

王文胜; 年诚旭; 张超; 阎如鹏; 吴鑫全; 张歆博

doi:10.13205/j.hjgc.202203024

基于YOLO v5模型的非住宅区自动垃圾分类箱设计

doi: 10.13205/j.hjgc.202203024

1. 北京信息科技大学机电工程学院, 北京 100192;
2. 北京信息科技大学自动化学院, 北京 100192

基金项目:

北京市科技计划项目(Z191100001419009)

详细信息

通讯作者:
王文胜(1990-),男,讲师,主要研究方向为计算机视觉、机械电子工程。ws_wang1128@126.com

计量
- 文章访问数: 263
- HTML全文浏览量: 60
- PDF下载量: 3
- 被引次数: 0
出版历程
- 收稿日期: 2021-09-25
- 网络出版日期: 2022-07-07

DESIGN OF AUTOMATIC GARBAGE SORTING BIN FOR NON-RESIDENTIAL AREA BASED ON YOLO v5

1. Mechanical and Electrical Engineering School, Beijing Information Science and Technology University, Beijing 100192, China;
2. School of Automation, Beijing Information Science and Technology University, Beijing 100192, China

摘要

摘要: 提出了一种基于YOLO v5模型的自动垃圾分类箱设计,应用于非住宅区的公共场所(如火车站、公交站、商场、校园等)。垃圾箱设计有4个垃圾桶,以2行2列摆放,中间为转轴,可带动轴上方的垃圾临时存储抽屉转动。采用单目摄像头采集视频图像,以英伟达Jetson nano嵌入式芯片作为上位机主控芯片,利用YOLO v5深度学习模型进行垃圾的自动提取与识别,并将上位机识别结果信息通过串口发送至下位机Arduino控制板,Arduino控制板控制舵机带动垃圾临时储存抽屉开口转动到相应的垃圾桶上方,从而控制升降台倾倒垃圾,完成垃圾自动分类。测试结果表明:垃圾识别结果稳定可靠,准确率可达到93%,能够实现垃圾自动分类。
- 垃圾分类 /
- YOLO v5 /
- 目标识别 /
- 深度学习
Abstract: A design of automatic garbage sorting bin based on model YOLO v5 was proposed, and applied to public places in non-residential communities (such as railway stations, bus stations, shopping malls, schoolyards, etc.). The trash bin was designed with 4 trash cans, arranged in two rows and two columns, with a rotating shaft in the middle, which drove the temporary storage drawer of garbage above the shaft to rotate. The monocular camera was used to collect video images, the embedded chip, Jetson nano by NVIDIA was used as the host computer's main control chip, and the YOLO v5 deep learning model was used for automatic garbage extraction and identification, and the recognition result information of the host computer was sent to the control board of the lower computer, Arduino, through the serial port. The control board, Arduino, controlled the motor to drive the opening of the temporary storage drawer to rotate to the top of the corresponding trash can, and then controlled the lifting platform to dump the trash and complete the automatic classification of the trash. The test results showed that the garbage identification results were stable and reliable, with an accuracy rate of 97%, thus the automatic garbage classification was realized.
- garbage classification /
- YOLO v5 /
- target recognition /
- deep learning

HTML全文

参考文献(21)

[1]	金佩薇,姚燕,梁晓瑜,等.垃圾图像识别研究进展[J/OL].环境工程:1-11. http://kns.cnki.net/kcms/detail/11.2097.X.20210621. 1039.006.html.
[2]	孙晓杰,王春莲,李倩,等.中国生活垃圾分类政策制度的发展演变历程[J].环境工程,2020,38(8):65-70.
[3]	彭韵,李蕾,彭绪亚,等.我国生活垃圾分类发展历程、障碍及对策[J].中国环境科学,2018,38(10):3874-3879.
[4]	董飞,扶漪红,吴笑天,等.城市生活垃圾分类治理:现实困境与实践进路[J].城市发展研究,2021,28(2):110-116.
[5]	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.
[6]	宋铁.基于机器视觉的家庭智能分类垃圾桶设计研究[D].上海:东华大学,2019.
[7]	马浚刚,朱振兴,杨梦龙,等.STM32F103C8T6的语音识别智能垃圾桶[J].电子世界,2021(14):104-106.
[8]	赵小芬.基于语音识别技术的垃圾分类收集系统研究[D].西安:陕西科技大学,2021.
[9]	王莉,何牧天,徐硕,等.基于YOLOv5s网络的垃圾分类和检测[J].包装工程,2021,42(8):50-56.
[10]	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.
[11]	陈亚宇,孙骥晟,李建龙,等.基于深度学习与图像处理的废弃物分类与定位方法[J].科学技术与工程,2021,21(21):8970-8975.
[12]	黄浩然.基于Hu不变矩的垃圾分类和识别[J].自动化应用,2020(8):74-76.
[13]	黄兴华,叶军一,熊杰.基于纹理特征融合的道路垃圾图像识别及提取[J].计算机工程与设计, 2019, 40(11):3212-3218 ,3305.
[14]	IRFAN S,BIMA S B D,IWAN K W. Visual-based trash detection and classification system for smart trash bin robot[C]//International Electronics Symposium on KnowLedge Creation and Intelligent Computing (IES-KCIC), 2018.
[15]	WANG Y, ZHANG X. Autonomous garbage detection for intelligent urban management[C]//MATEC Web of Conferences. EDP Sciences, 2018, 232:01056.
[16]	KUMAR S, YADAV D, GUPTA H, et al. A Novel YOLOv3 Algorithm-Based Deep Learning Approach for Waste Segregation:Towards Smart Waste Management[J]. Electronics, 2021, 10(1):14.
[17]	许伟,熊卫华,姚杰,等.基于改进YOLOv3算法在垃圾检测上的应用[J].光电子·激光, 2020, 31(9):928-937.
[18]	赵冬娥,吴瑞,赵宝国,陈媛媛.高光谱成像的垃圾分类识别研究[J].光谱学与光谱分析,2019,39(3):921-926.
[19]	张珂,冯晓晗,郭玉荣,等.图像分类的深度卷积神经网络模型综述[J].中国图象图形学报,2021,26(10):2305-2325.
[20]	康庄,杨杰,郭濠奇.基于机器视觉的垃圾自动分类系统设计[J].浙江大学学报(工学版),2020,54(7):1272-1280,1307.
[21]	WANG W F, ZHANG B B, WANG Z Q, et al. Garbage image recognition and classification based on HOG feature and SVM-Boosting[J]. Journal of Physics Conference Series, 2021, 1966(1):012002.