作者：汪洋，王小妮，王育新，刘畅，熊继伟，韩定良
单位：北京信息科技大学理学院，北京 100101
中图分类号：TP242.6              
文献标识码：A              
文章编号：1006-883X(2020)08-0019-07
收稿日期：2020-06-29

摘要：近些年来，我国各地陆续出台了垃圾分类的政策，垃圾分类也已经逐步成为了一种新的生活方式。但是垃圾分类又存在着分类效率低、分类成本高等问题。针对这些问题提出了一种基于卷积神经网络的垃圾分类系统。通过软硬件相结合的方式实现了垃圾投放检测、垃圾种类识别、垃圾精确投放、结果反馈等功能。对于日常生活垃圾的识别率已达91.7%以上，具有识别率高、分类速度快、方便迭代更新、成本低等优点。
关键词：垃圾分类；机器视觉；卷积神经网络；深度学习   

Research on Garbage Classification System Based on Convolution Neural Network
WANG Yang, WANG Xiao-ni, WANG Yu-xin, LIU Chang  XIONG Ji-wang, HAN Ding-liang
School of Science, Beijing Information Science and Technology University, Beijing 100192, China
Abstract: In recent years, garbage classification policies have been introduced in China, and garbage classification has become a new way of life. But there are some problems in waste classification, such as low efficiency and high cost. To solve these problems, we propose an intelligent garbage classification system based on convolutional neural network. Through the combination of software and hardware, the functions of garbage detection, garbage type identification, garbage precise delivery and result feedback are realized. The recognition rate of MSW is more than 91.7%. The system has the advantages of high recognition rate, fast classification speed, convenient iterative update and low cost.
Key words: garbage classification; machine vision; convolutional neural network; deep learning  

阅读全文

备注：2020年 第26卷 第08期