Design and Implementation of a Raspberry Pi-Based Smart Wet and Dry Waste Segregation System Using AI

Abstract

— With increasing urbanization and consumerism, the global Municipal Solid Waste (MSW) generation has skyrocketed. The ever-growing MSW collection, transportation and disposal infrastructure is insufficient and is overwhelmed by the volume of waste generated. Manual segregation of wet and dry MSW is a cumbersome process because of unmethodized sorting, low accuracy and 60-70% of wet recyclable waste still head into landfills. We hence propose and analyze an AIenabled intelligent, automated, and low-cost trash segregation prototype that classifies household level wet or dry trash using camera modules and onboard computer module, runs convolution neural network model on raspberry pi 4 to detect the object, actuates servo pan-tilt into the bin, and loads onto ultrasonic sensors to monitor bin fill level and load cells with HX711 amplifier to measure the dry and wet mass accumulated in each bin. Such an intelligent waste segregation system can cater to home, school, Institutional and community level waste segregation needs.