基于深度学习和Arduino的生活垃圾分类装置设计

打开文本图片集

中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2025）18-0179-05

Abstract： Aiming at the problems oftraditional household waste classification methods relying on manual work，low eficiencyandhighclasificationerrorrate，ahousehold wasteclasificationdevicebasedonDeepLearmingandArduinois designed.Thedevicetrainsandoptimizes the waste imagerecognition model throughDeep Learningalgorithm，deploys the modelonK210，andconnects thevoicebroadcastmodule，servo motorcontrolmodule，temperatureandhumiditymeasurement moduleandfulloaddetectionmodulethroughArduinodevelopmentboard，soastorealizethewasteclassficationfunction. The experimental results show thattherecognition accuracyof the device for four diferent kinds of waste is more than 90% （20 under three different light intensities. In normal use，the average recognition rate is as high as 97% .At the same time， the voice broadcast function，temperatureand humidity detection function and fulload detectioncan operate normallyduring the test process，which provides an intelligent solution for household and community domestic waste clasification.

Keywords：Deep Learning;Waste classification; Arduino; image classification

0 引言

“十四五”以来，习近平总书记重点强调“考察必看生态，坚持绿色发展”的指导思想，垃圾分类回收是我国高质量绿色发展战略中的重要环节。（剩余6802字）