Automatic Segregation of Limited Wastes through Tiny YOLOv3 Algorithm
Neil Alexander G. Macasaet, Elidad Rachel R. Martinez, Ernesto M. Vergara · 2022 13th International Conference on Computing Communication and Networking Technologies (ICCCNT) · 2022
To maintain a clean environment, proper waste management must be put into effect. This maintains a safe and healthy environment by preventing surface waterways and land regions contamination, reducing air pollution, and suppressing flooding and disease outbreaks. Using innovative technology can effectively help enhance waste management. In this paper, the researchers developed a device that segregates limited wastes through the application of tiny YOLOv3. The researchers were able to detect, classify, and segregate limited biodegradable wastes from non-biodegradable wastes using a pre-trained model of Tiny YOLOv3 on a Raspberry Pi 4B. Based on the collected data, the system produced macro-averaged precision, recall, and F1 score values for detection and classification of 100%, 88%, and 93%, respectively. Additionally, it had a detection and classification accuracy of 88% overall and a waste segregation accuracy of 100%.