A smart municipal waste management system based on deep-learning and Internet of Things
Cong Wang,Jiongming Qin,Cheng Qu,Xu Ran,Chuanjun Liu,Bin Chen
DOI: https://doi.org/10.1016/j.wasman.2021.08.028
Abstract:A proof-of-concept municipal waste management system was proposed to reduce the cost of waste classification, monitoring and collection. In this system, we utilize the deep learning-based classifier and cloud computing technique to realize high accuracy waste classification at the beginning of garbage collection. To facilitate the subsequent waste disposal, we subdivide recyclable waste into plastic, glass, paper or cardboard, metal, fabric and the other recyclable waste, a total of six categories. Deep-learning convolution neural networks (CNN) were applied to realize the garbage classification task. Here, we investigate seven state-of-the-art CNNs and data pre-processing methods for waste classification, whose accuracies of nine categories range from 91.9 to 94.6% in the validation set. Among these networks, MobileNetV3 has a high classification accuracy (94.26%), a small storage size (49.5 MB) and the shortest running time (261.7 ms). Moreover, the Internet of Things (IoT) devices which implement information exchange between waste containers and waste management center are designed to monitor the overall amount of waste produced in this area and the operating state of any waste container via a set of sensors. According to monitoring information, the waste management center can schedule adaptive equipment deployment and maintenance, waste collection and vehicle routing plans, which serves as an essential part of a successful municipal waste management system.