Aplikasi Sistem Deteksi Sampah Organik dan Non Organik Menggunakan Algoritma YOLO V8

Abstract

Effective waste management is a crucial global challenge, especially in addressing environmental and public health issues. In Indonesia, increasing the volume of organic and non-organic waste requires innovative solutions for more efficient waste separation. This research aims to develop an application of an automatic waste detection system at Temporary Disposal Sites (TPS) using the You Only Look Once algorithm version 8 (YOLO v8), the latest object detection algorithm that offers high speed and accuracy. The first stage will be carried out this field study to collect a wide sample of waste datasets, including various types of organic and non-organic waste, to train the YOLO v8 model. The second stage is to build a computational model, this model is built based on the you only look once (YOLO) method version 8 and the third stage is to build a system that suits the needs of the field and test the system against the needs. This study shows that YOLOv8 has excellent performance in detecting organic and non-organic waste, model evaluation with an average accuracy of 99.35%, precision of 98.6%, recall of 98.6%, and f1-score of 98.5%. These results show that the YOLOv8 method can speed up and simplify the waste sorting process, so that it can be used and used by the community in waste disposal sites both in the surrounding home environment and public places to automatically detect organic and non-organic waste categories.