针对生活垃圾的智能分类垃圾桶设计与制作
Design and Production of the Intelligent Classification Trash Can for Domestic Garbage
杜雪林 1邹家华 1刘颖 1陈毅 1涂宇 2易文慧 1王港1
作者信息
- 1. 湖南工学院 智能制造与机械工程学院,湖南衡阳 421002
- 2. 湖南交通工程学院 交通运输工程学院,湖南衡阳 421002
- 折叠
摘要
针对智能分类垃圾处理装置研制的关键问题,包括视觉识别、机构设计和满载检测等,研发了一种新型智能分类垃圾桶.该垃圾桶具有开机自动循环播放宣传视频、自动垃圾识别和分类、可回收垃圾压缩、餐厨垃圾粉碎和满载报警等功能.为提高垃圾识别准确度,采用神经网络MobileNetV2模型和迁移学习方法,基于自建垃圾数据集进行了训练和测试.通过树莓派4B开发板实现分类机构、粉碎机构和压缩机构的运动控制.在制作的智能分类垃圾桶实物样机上进行系统测试,验证了各相关功能的可靠性,为智能分类垃圾处理装置的研发提供参考.
Abstract
A new type of intelligent classification trash can has been developed to address key issues in the development of intelligent classification trash treatment devices,including visual recognition,mechanism design and full load detection.This trash can has functions such as automatic playback of promotional videos upon startup,automatic garbage recognition and classification,compression of recyclable garbage,crushing of kitchen waste,and full load alarm.To improve the accuracy of garbage recognition,a neural network MobileNetV2 model and transfer learning method are used for training and testing based on a self-built garbage dataset.The motion control of the classification mechanism,crushing mechanism and compression mechanism is achieved through the Raspberry Pi 4B development board.A system test is conducted on the physical prototype of the intelligent classification trash can,which verifies the reliability of various related functions and provides a reference for the development of intelligent classification trash treatment devices.
关键词
智能分类垃圾桶/MobileNetV2模型/迁移学习/粉碎机构/压缩机构Key words
intelligent classification trash can/MobileNetV2 model/transfer learning method/crushing mechanism/compression mechanism引用本文复制引用
出版年
2024
NETL
NSTL
万方数据