基于YOLOv5的智能垃圾分类系统设计

Waste Sorting and Detection Based on YOLOv5

目前我国正在大力推广垃圾分类制度,垃圾分类制度不仅与国民生活环境息息相关也和我国生态文明建设紧密关联。当前,我国垃圾分类方式效率有限,无法更好地利用垃圾资源,为了提高垃圾分类效率和资源利用率,本文提出了基于YOLOv5的智能垃圾分类系统,采用STM32单片机作为控制器,使用外接的摄像头进行图像采集,通过YOLOv5模型进行图像识别,将识别结果通过串口通信发送给单片机来控制步进电机,进行垃圾投放,从而实现了垃圾自动识别和分拣的功能。该系统将深度学习与嵌入式开发应用到垃圾分类问题中,经过测试,能达到较好的识别率且操作简单,可复制性强。

At present,China is vigorously promoting the garbage classification system,which is related to the national living environment and the construction of China's ecological civilization.At present,China's garbage classification method has limited efficiency,can not better use garbage resources,in order to improve garbage classification efficiency and resource utilization,this paper proposes an intelligent garbage classification system based on YOLOv5,using STM32 single-chip microcomputer as a controller,using an external camera for image acquisition,image recognition through YOLOv5 model,the recognition results are sent to the single-chip microcomputer through serial communication to control the stepper motor,and garbage is put in,so as to realize the function of automatic garbage identification and sorting.The system applies deep learning and embedded development to garbage classification problems,and after testing,it can achieve a good recognition rate,simple operation and strong reproducibility.