基于卷积神经网络的智能垃圾分拣车  

Intelligent Garbage Sorting Vehicle Based on Convolutional Neural Network

作　　者：石则斌 罗雪峰 王隐 闻春敖[1] 蔡佩君[1] SHI Zebin;LUO Xuefeng;WANG Yin;WEN Chunao;CAI Peijun(College of Optoelectronic Science and Engineering,Zhejiang University,Hangzhou 310027,China)

机构地区：[1]浙江大学光电科学与工程学院,杭州310027

出　　处：《实验室研究与探索》2022年第12期123-126,共4页Research and Exploration In Laboratory

摘　　要：基于卷积神经网络,设计一种智能垃圾分拣车。该分拣车以树莓派为控制核心,使用单目广角摄像头、超声波测距传感器、倾斜角度传感器获取信息,以麦克纳姆轮小车为运动载体,通过Yolo-Fastest目标检测算法和自主编写的运动反馈控制算法,实现自主路径规划、垃圾识别、垃圾抓取、垃圾分类等功能,具有简单可靠、性价比高的特点。An intelligent garbage sorting vehicle is designed based on convolutional neural network.This vehicle takes Raspberry Pi as the control core,uses monocular wide-angle camera,ultrasonic ranging sensor and tilting angle sensor to obtain information,takes Mecanum wheel car as the motion carrier,and realizes the functions of autonomous path planning,garbage identification,garbage garbing and garbage classification through Yolo-Fastest target detection algorithm and self-written motion-feedback-control algorithm.It is simple,reliable and cost-effective,and contributes to intelligent refuse classification.

关 键 词：卷积神经网络 垃圾分类 智能垃圾分拣车 树莓派 目标检测 

分 类 号：S37[农业科学—农产品加工]