A sustainable approach to waterlogging prevention: Solar-powered garbage detection using machine learning

Abstract

Waterlogging is a persistent problem in Dhaka, especially during the rainy season which is mostly caused by the poor waste management system that blocks the sewers. Along with imposing serious public health risks, waterlogging disrupts daily life largely and puts a huge negative impact on the overall economy. In this paper, we propose a smart pole system configured to prevent waterlogging by effective waste identification and disposal to solve this issue. The system includes a 360° camera, a dual-axis solar panel for sustainable power generation, a circuit box to manage the energy supply and automated LED lighting for night illumination. Moreover, machine learning and YOLOv8s model is used for faultless roadside garbage detection. MATLAB is used to perform solar panel simulation, ensuring uninterrupted function by optimizing energy produced by the solar panel. The key goal of this system is to identify trash and autonomously alert the city corporation to remove the waste and avoid drain clog which leads to waterlogging. Finally, implementing this system across the city could offer a feasible and sustainable solution for urban waste management and waterlogging problems.