Skripsi/Tugas Akhir
Deteksi Objek Sampah Botol Plastik dengan Deep Learning Menggunakan Model You Only Look Once (YOLO)
ABSTRAK
Permasalahan limbah plastik, terutama botol plastik, semakin menjadi perhatian serius dalam isu lingkungan. Deteksi otomatis terhadap sampah botol plastik dapat meningkatkan efisiensi dalam pengelolaannya. Penelitian ini berfokus pada pengembangan model deteksi objek botol plastik berbasis deep learning dengan menerapkan algoritma You Only Look Once (YOLO). Metode yang diterapkan mencakup pengumpulan dan pelabelan data, pelatihan model menggunakan YOLO, serta evaluasi kinerjanya melalui confusion matrix dan metrik evaluasi lainnya. Hasil evaluasi menunjukkan bahwa model mampu mendeteksi botol plastik dengan baik. sampah botol plastik. Berdasarkan hasil confusion matrix, model memperoleh akurasi sebesar 92% dalam mengidentifikasi botol plastik dan 82% untuk kategori bukan botol plastik. Selain itu, model memiliki nilai precision sebesar 95,53%, recall 88,76%, mean Average Precision pada threshold 0,5 (mAP50) sebesar 92,60%, serta mean Average Precision pada berbagai threshold 0,5–0,95 (mAP50- 95) sebesar 76,87%. Dengan hasil tersebut, model ini diharapkan dapat berkontribusi dalam sistem pengelolaan sampah berbasis teknologi guna mendukung proses identifikasi dan klasifikasi sampah secara otomatis.
Kata Kunci: Deteksi Objek, Sampah Botol Plastik, Deep Learning, YOLO
ABSTRACT
The issue of plastic waste, particularly plastic bottles, has become an increasingly serious environmental concern. Automatic detection of plastic bottle waste can enhance efficiency in its management. This research focuses on developing an object detection model for plastic bottles based on deep learning using the You Only Look Once (YOLO) algorithm. The applied method includes data collection and labeling, model training using YOLO, and performance evaluation through a confusion matrix and other evaluation metrics. The evaluation results indicate that the model performs well in detecting plastic bottles. Based on the confusion matrix results, the model achieved an accuracy of 92% in identifying plastic bottles and 82% for the non-plastic bottle category. Additionally, the model obtained a precision of 95.53%, recall of 88.76%, mean Average Precision at a 0.5 threshold (mAP50) of 92.60%, and mean Average Precision across various thresholds from 0.5 to 0.95 (mAP50-95) of 76.87%. With these results, this model is expected to contribute to technology-based waste management systems, supporting the automated identification and classification of waste.
Keywords: Object Detection, Plastic Bottle Waste, Deep Learning, YOLO
Tidak ada salinan data
Universitas DIPA Makassar
NPP 7371142D1000002
Jln. Perintis Kemerdekaan KM.9
Telp. (0411)587194
Hotline: +6281228221994
WhatsApp Admin: +6281342092072
e-Mail: [email protected]
© 2025 — Perpustakaan UNDIPA Makassar - SLiMS