Publication

Abstract : Automated waste segregation is crucial for efficient and sustainable waste management in modern residences. This paper details the conceptualization, development, and performance assessment of a smart waste classification system integrating visual recognition with microcontroller-driven automation. A centralized unit, driven by a microcontroller and webcam, uses machine vision and learning to sort waste into biodegradable and non-biodegradable groups. The system's smart core uses a powerful neural network, running on special hardware, to look at and sort images of waste. A microcontroller platform manages the physical sorting process, controlling a motor-powered conveyor belt and servo motors that move waste into the right bins depending on its classification. Infrared sensors help track the movement and position of waste to ensure accurate sorting. The system's performance was tested based on how well it classifies waste, how efficiently it sorts, and how fast it processes. The results show that using a complex deep learning model for precise waste sorting is possible with a cost-effective microcontroller for real-time control of mechanical systems. This adaptable and scalable method helps automate waste sorting, lowers the need for manual work, improves recycling, and supports eco-friendly habits in smart homes. It shows how combining advanced image analysis with strong microcontroller technology can lead to better waste management.