Computer-vision-powered Automatic Waste Sorting Bin: a Machine Learning-based Solution on Waste Management

Abstract The mainstream waste separation methods – recycle bins and manual are time consuming, prone to error, human labor intensive. Robots have demonstrated their superior efficiency at repetitive task over human’s again. This research implements evaluate a computer-vision-powered Automatic Waste Sorting Bin, which is capable classify the types in short with high efficiency. Controlled classified using Raspberry pi camera, bin can detect type drop it right accordingly. dataset on model trained relatively small. An image of first captured then analyzed YOLOv5 model. Parameters that yield optimal result 150 epochs YOLOv5l an accuracy 93.33 %. To improve model’s performance, we experimented different epoch settings measured results. setup proposed this paper provides automated solution replace separation. low-cost flexible be easily replicated larger acquire better result. Future may reference as proof concept and/or implementation for streamlined solution.