|
Efficient Waste Detection and Classification based on YOLOv5 Models |
|---|---|
| รหัสดีโอไอ | |
| Creator | Suksun Promboonruang |
| Title | Efficient Waste Detection and Classification based on YOLOv5 Models |
| Contributor | Thummarat Boonrod, Yongyut Ratchatawetchakul |
| Publisher | Faculty of Engineering Mahasasakham University |
| Publication Year | 2567 |
| Journal Title | Engineering Access |
| Journal Vol. | 10 |
| Journal No. | 1 |
| Page no. | 51-58 |
| Keyword | YOLOv5, Waste Detection, Artificial Intelligence, Image Processing, Efficient Waste Detection |
| URL Website | https://ph02.tci-thaijo.org/index.php/mijet/index |
| Website title | THAIJO Engineering Access |
| ISSN | 2730-4175 |
| Abstract | This paper proposes efficient waste detection and classification based on YOLOv5 by utilizing YOLOv5 for waste detection and classification. Divide the dataset into 4 classes consisting of wood, glass, plastic, and metal. The dataset is methodically divided into three subsets: the training set consisting of 1,860 images, the validation set consisting of 200 images, and the test set consisting of 235 images. The objective of our study is to assess the effectiveness of three YOLOv5 models, namely Yolov5s, Yolov5m, and Yolov5x, across several waste object categories. The methodology employed in this research is as follows: Compilation of datasets and development of models specific to each iteration of YOLOv5. Comparing models. We assess the precision, recall, and mean average precision (mAP) to measure the correctness and speed of their processing. The empirical findings from our investigation suggest that Yolov5x demonstrates the highest level of accuracy and mAP scores (0.41), whilst Yolov5s showcases the shortest processing time (0.83 hours). |
