摘 要:由于人们生活水平的不断提高,随之产生的生活垃圾也不断增多。为解决靠人力进行垃圾分类效率低下的问题,对自动垃圾分拣系统进行研究,并设计一种基于机器视觉的垃圾分拣系统。使用搭载有摄像头的树莓派作为主控单片机,利用卷积神经网络来分辨摄像头所拍摄垃圾的类型,借助 Arduino 控制舵机自动开合垃圾桶,实现垃圾分类。经过测试分析,系统的准确精度均能达到 70% 以上,准确精度较好,籍此人们能够轻松处理生活垃圾。
关键词:机器视觉;树莓派;卷积神经网络;Arduino
DOI:10.19850/j.cnki.2096-4706.2023.04.005
基金项目:校一般科研课题(njpj2021-2-04)
中图分类号:TP311 文献标识码:A 文章编号:2096-4706(2023)04-0018-05
Design of Garbage Sorting System Based on Machine Vision
CHEN Meiling1, ZHU Wenhan2, LIU Jiacheng1, ZHENG Yang3
(1.Nanjing Tech University Pujiang Institute, Nanjing 211200, China; 2.Graduate School Qinhuangdao Branch, Northeastern University, Qinhuangdao 066004, China; 3.Jiamusi University, Jiamusi 154007, China)
Abstract: With the continuous improvement of people's living standards, the amount of domestic garbage generated is also increasing. In order to solve the problem of inefficient waste sorting by manpower, the automatic waste sorting system is studied and a waste sorting system based on machine vision is designed. The Raspberry pie equipped with a camera is used as the main control MCU, and the Convolution Neural Network is used to distinguish the type of garbage captured by the camera. The steering gear is controlled by Arduino to automatically open and close the garbage can, realizing garbage classification. After testing and analysis, the accuracy of the system can reach more than 70%, and the accuracy is good, so people can easily handle domestic garbage.
Keywords: machine vision; Raspberry pie; Convolution Neural Network; Arduino
参考文献:
[1] 王蓉蓉 . 基于机器学习的垃圾识别与分类研究 [D]. 兰州:兰州理工大学,2021.
[2] 李健凯 . 基于深度学习的视觉机器人垃圾分拣系统设计[D]. 天津:天津工业大学,2021.
[3] 肖扬 . 基于机器视觉的废纸箱分类识别技术研究 [D]. 郑州:河南大学,2020.
[4] 王喆 . 基于机器视觉的医疗垃圾桶识别抓取方法研究 [D].济南:齐鲁工业大学,2020.
[5] 张凤 . 智能垃圾分拣系统关键技术研究 [D]. 成都:成都理工大学,2020.
[6] 黄帅 . 基于机器视觉的大型生活垃圾焚烧过程诊断方法研究 [D]. 杭州:浙江大学,2020.
[7] 段昆昕,江秋仪,陈锋楠,等 . 基于树莓派下智能分类垃圾桶设计 [J]. 中国新技术新产品,2022(1):37-39.
作者简介:陈美玲(1984—),女,满族,吉林四平人,副教授,硕士,研究方向:智能控制、图像处理。