SPubBin: Smart Public Bin Based on Deep Learning Waste Classification: An IOT system for Smart Environment in Algeria

Abstract

Over the past few years, Internet of things (IoT) has become one of the most important technologies of the 21st century. Nowadays, it is possible to connect different objects to the internet via embedded devices. By means of this technology, physical objects can share and collect data with minimal human contribution. In this inter-connected world, digital systems can record, monitor, and adjust each interaction between connected things. The IoT technologies support smart cities initiatives around the world in order to promote greener and safer urban environments, with cleaner air and water, and better public services and mobility. Smart environment is an inseparable part of smart cities. In our environment, wastes are objects that we put in the Bin. Waste management sorting and recycling are two central elements in the fight against climate change. Existing solid waste management policies and methods are being challenged by the exponential population expansion in urban areas. In fast-growing cities, where increasing garbage output exceeds the capacity of current facilities, the issue becomes even more difficult. Waste avoidance, recycling, reuse, and recovery are essential factors for reducing solid waste discharged in landfills, particularly in rapidly increasing cities where more sustainable management techniques are required. In smart home, effective household waste management is essential to building habitable cities however remains a challenge for many developing countries and cities. In this context, as any other important aspect of city management, a good waste disposal strategy plays a crucial role for making cities more glamorous. In this paper, a novel approach for waste sorting is proposed. Baptized SPubBin (Smart Public Bin), the presented solution is based on transfer learning, and uses, three CNNs models: VGG16, Dense201 and Resnet50. In order to validate the proposed approach, we have developed a tool supporting it.