YOLOv8-based Waste Detection System for Recycling Plants: A Deep Learning Approach

Waste management is increasingly attracting attention due to its role in smart and sustainable development, especially in developed and developing nations. This system consists of a series of interconnected processes that perform various complex functions. Deep learning has recently attracted interest as an alternative computational method to solve various waste classification challenges. Many researchers have focused on this area, yielding significant research results in recent years. Although several in-depth investigations have been conducted on waste detection and classification, the WaRP dataset was created specifically to train and evaluate the proposed algorithms using industrial data from conveyor belt of waste recycling plant. Surprisingly, no research has explored the application of the YOLOv8 model to solve the waste management problem using the WaRP dataset. This experiment makes a notable contribution by detecting waste through a pyramid and direct prediction method, which differs from the traditional model based on anchor boxes. Through experimentation with different YOLOv8 model weights, this research study found that YOLOv8s provides relatively good results with smaller dataset and lower processing time. On the other hand, YOLOv81 achieves a higher mAP50 value of about 59% on the same dataset, but at the cost of high inference time.