基于卷积神经网络的垃圾分类小车系统  

Garbage Sorting Trolley System Based on Convolutional Neural Network

作　　者：孙家凡 严培 魏瑜晗 万伟豪 李国豪 SUN Jiafan;YAN Pei;WEI Yuhan;WAN Weihao;LI Guohao(Nanjing Institute of Technology,Nanjing,Jiangsu 211167,China)

机构地区：[1]南京工程学院,江苏南京211167

出　　处：《自动化应用》2023年第7期64-66,共3页Automation Application

基　　金：江苏省高等学校大学生实践创新训练项目(202211276050Z)。

摘　　要：目前,垃圾分类主要靠民众自主分类,以校区或居住点为单位,将已经初步分类的垃圾运输后处理,此方式往往容易导致错漏,效率低,针对该问题,本文提出基于神经网络的垃圾分类小车系统,旨在通过机械化的标准规范化初步垃圾分类,简化垃圾分类问题,提高资源再利用率。该系统通过基于卷积神经网络的YOLO物体识别系统识别垃圾种类后,利用机械臂自主夹取,以完成垃圾分类。此外,小车配备自主避障系统,能适应一般环境的路线规划,为垃圾分类的普及化和简易化提供思路。The current approach to garbage classification depends largely on the autonomous efforts of the public,where garbage is sorted at the campus or residential level and then transported for processing.Unfortunately,this method is prone to errors,leakage,and inefficiencies.To overcome this challenge,this paper proposes a garbage classification trolley system based on a neural network.The goal of this system is to standardize the initial garbage sorting process through mechanization and improve the efficiency of garbage management.This approach utilizes the YOLO object recognition system,which relies on a convolutional neural network,to identify the different types of garbage.The identified garbage is then automatically picked up by a robot arm for accurate sorting.Furthermore,the system includes an autonomous obstacle avoidance system that can adapt to different environments,improving its route planning capabilities.By simplifying the garbage classification process,this approach can provide valuable insights for the wider adoption of efficient and standardized waste disposal practices.

关 键 词：垃圾分类 卷积神经网络 YOLO物体识别 智能导航避障 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]