Title: Waste plastic bottles classification with deep learning model

Authors: Jixu Hou; Xiaofeng Xie; Wenwen Wang; Qian Cai; Zhengjie Deng; Houqun Yang; Hongnian Huang; Yizhen Wang

Addresses: College of Physics & Electronic Engineering, Hainan Normal University, Haikou, Hainan, China ' College of Mechanical & Electronic Engineering, Hainan University, Haikou, Hainan, China ' College of Foreign Languages, Hainan Normal University, Haikou, Hainan, China ' College of Mechanical & Electronic Engineering, Hainan Normal University, Haikou, Hainan, China ' College of Mechanical & Electronic Engineering, Hainan Normal University, Haikou, Hainan, China ' College of CSCI & Technology, Hainan University, Haikou, Hainan, China ' Department of Mathematics and Statistics, University of New Mexico, Albuquerque, New Mexico, USA ' College of Physics & Electronic Engineering, Hainan Normal University, Haikou, Hainan, China

Abstract: The misuse of plastic products has led to serious environment problem. To alleviate such phenomenon, we need to recover the plastic waste with a precise distinction. In this work, we applied a deep learning model, e.g., Faster-RCNN, to identify the class of plastic bottle. We have designed a waste plastic bottle recycling system, which can cooperate with the manipulator and conveyor to automatically sort the bottles in the garbage. During the experiment, we established a data set containing 8400 images. Different backbone networks are used to train on the data set. The experimental results show that the skeleton network using Resnet-50 as Faster-RCNN has higher detection performance than other networks. The system can also be applied to the identification and classification of other solid wastes.

Keywords: plastic pollution; object detection; garbage classification; plastic bottle recycling; automatic sorting.

DOI: 10.1504/IJWMC.2023.134675

International Journal of Wireless and Mobile Computing, 2023 Vol.25 No.3, pp.296 - 302

Received: 29 Oct 2021
Received in revised form: 18 Apr 2022
Accepted: 30 May 2022
Published online: 03 Nov 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article