Waste Sorting as a Service is an innovative deep learning-based waste management system smart technology for advanced waste classification. The system’s core consists of a Convolutional Neural Network (CNN) where a voluminous image database is trained to sort and identify waste, where wastes are identified under different classifications including organic, plastics, glass, and paper. Once the system recognizes through image recognition the type of waste it labels according to the context, the motors powered by Arduino open the proper bin compartment for sorting in real-time. Where mixed or incorrectly classified waste is found, the system forbids the opening of any bin to allow users to dispose of the waste correctly. Each part of the smart bins is designed to accommodate particular sorts of waste, which lessens the intermingling of waste sorts and raises the recycling rate. The combination of CNN-based image recognition applied to bin identification with the automatic control of the bins, on the other hand, is not only efficient and convenient in handling the waste disposal system but also beneficial in helping the cause of environmentalism through a decrease in the volume of waste ending up in the landfills as well as encouraging everyone and anybody into adopting the correct ways of recycling. The design of the system allows the extension to both urban environments and concepts of smart cities, the solutions for granting sustainable waste management and at the same time using up-to-date automation and deep learning technologies in real-time.
Keywords: Arduino-controlled smart bin, Live classification, Waste Management, deep learning
