基于機器視覺的垃圾分類系統(tǒng)設計

打開文本圖片集

關鍵詞：YOL0V5模型；垃圾分類；深度學習

doi：10.3969/J.ISSN.1672-7274.2025.05.022

中圖分類號：TP31 文獻標志碼：A 文章編碼：1672-7274（2025）05-0064-03

Abstract： This paper realizes automatic identification of garbage images through the YOLOV5s model， introduces the CBAM attention mechanism to optimize the performance of the YOLOV5s model in practical applications.By equipping an STM32 microcontroller as the main control device，the stepper motor is controlled to sort garbage based on the clasificationresults oftheYOLOV5s model.TVOCand infraredsensing modules are integrated to monitor the internalenvironmentof thetrash can inreal-time.This design applies deep learning modelsand embeddeddevices to garbage classification， which holds significant practical value for waste sorting.

Keywords： YOLOV5s model; waste sorting; deep learning

0 引言

隨著經(jīng)濟的發(fā)展，生活垃圾運量呈現(xiàn)上升趨勢，截至2022年，我國生活垃圾運量高達24444.7萬噸，垃圾分類平均覆蓋率達到 82.5% ，但目前城市的生活垃圾分類存在分類效率低下，人力耗費較高，處理環(huán)境惡劣等問題。（剩余3628字）