摘 要:目前,人类对资源需求的日益增加造成严重的环境问题,如何有效地实施环境保护是人类共同面临的课题。按照垃圾分类准则,文章设计制作了一个 2 000 张的垃圾数据集,并分为四类。利用 ResNet 进行目标分类网络模型训练,将训练好的模型载入到主控模块,结合 HuskyLens 视觉传感器、舵机分拣模块、红外感应溢满模块,实现对目标垃圾的检测识别、自动分拣及溢满报警。该设计完成度高,稳定性较强,后期可进行工程实现投入实际生活中。
关键词:神经网络;垃圾分类;HuskyLens 视觉传感器;目标检测
DOI:10.19850/j.cnki.2096-4706.2022.17.011
基金项目: 中原科技创新领军人才(214200510013);河南省高校重点科研项目(21A510016,21A520052);留学人员科研资助和创业启动项目(HRSS2021[36]);校内重大项目成果培育计划(K2020ZDPY02)
中图分类号:TP18 文献标识码:A 文章编号:2096-4706(2022)17-0045-04
Design of Intelligent Garbage Sorting System Based on Deep Learning
LI Pingyuan1, SONG Xiaowei 2, WANG Yuying1
(1.School of Electronic and Information Engineering, Zhongyuan University of Technology, Zhengzhou 450007, China; 2.Kaifeng University, Kaifeng 475004, China)
Abstract: At present, the increasing human demand for resources has caused serious environmental problems. How to effectively implement environmental protection is a common topic faced by mankind. According to the garbage classification criteria, this paper designs and makes a 2 000 garbage data set, which is divided into four categories. ResNet is used to train the target classification network model, and the trained model is loaded into the main control module. Combined with HuskyLens visual sensor, steering gear sorting module and infrared sensing overflow module, the detection and identification of target garbage, automatic sorting and overflow alarm are realized. The design has high degree of completion and strong stability, and it can be realized for project and put into practical life in the later stage.
Keywords: neural network; garbage sorting; HuskyLens visual sensor; target detection
参考文献:
[1] 卢淑怡,魏爽,万思远 . 基于深度学习的智能分类垃圾桶[J]. 计算机与数字工程,2021,49(5):1009-1012+1029.
[2] 吕文杰,魏孝虎,陈众孚,等 . 基于卷积神经网络的生活垃圾自动分类软件的实现 [J]. 电脑知识与技术,2020,16(5):203-204.
[3] 吴碧程,邓祥恩,张子憧,等 . 基于卷积神经网络的智能垃圾分类系统 [J]. 物理实验,2019,39(11):44-49.
[4] MILIOS L,DAVANI A E,YU Y.Sustainability impact assessment of increased plastic recycling and future pathways of plastic waste management in Sweden [J].Recycling,2018,3(3):33-33.
[5] 孟小燕,王毅,苏利阳,程多威,郝亮 . 我国普遍推行垃圾分类制度面临的问题与对策分析 [J]. 生态经济,2019,35(5):184-188.
作者简介:李平原(1995—),男,汉族,河南信阳人,硕士研究生在读,主要研究方向:计算机视觉;宋晓炜(1978—),男,汉族,山西大同人,教授,博士,主要研究方向:图像处理、计算机视觉;王玉莹(1998—),女,汉族,河北三河人,硕士研究生在读,主要研究方向:计算机视觉。