ABSTRAKPermukaan kayu mengalami berbagai serangan serangga dan jamur, sehingga dapat menyebabkan cacat seperti pembusukan pada kayu yang berpengaruh terhadap kualitas dan harga jual kayu tersebut. Pengujian secara lapangan dengan penglihatan manusia menjadi kurang efektif, karena menghasilkan penilaian yang subjektif dan memerlukan waktu yang lama. Penelitian ini mengusulkan sistem deteksi cacat pada permukaan kayu pinus dan kayu karet menggunakan Convolutional Neural Network (CNN) dengan model YOLOv5. Pengujian sistem dilakukan menggunakan beberapa model dari YOLOv5, serta dua teknik image enhancement, yaitu edge filter dan Real ESRGAN. Hasil mAP terbaik sebesar 94,3% dengan kecepatan 125 FPS pada dataset kayu pinus menggunakan model YOLOv5s tanpa penambahan image enhancement. Sedangkan pada dataset kayu karet yang memiliki jenis cacat yang lebih kompleks, hasil mAP terbaik adalah sebesar 94,7% dengan kecepatan 139 FPS menggunakan model YOLOv5s-Transformer dengan penambahan image enhancement Real ESRGAN.Kata kunci: deteksi, kayu, Convolutional Neural Network (CNN), YOLO ABSTRACTWood surafce is subject to various insect and fungal attacks, which can cause defects such as wood rot. This condition affects the quality as well as the selling price of the wood. Field testing with human eyesight becomes less effective because it produces a subjective assessment and time consuming. This study proposes a defect detection system on the surface of pine wood and rubber wood using the Convolutional Neural Network (CNN) with the YOLOv5 model. System testing was carried out using several models from YOLOv5 and two image enhancement techniques, namely edge filter and Real ESRGAN. The best mAP results were 94.3%, with a speed of 125 FPS on the pine wood dataset using the YOLOv5s model without adding image enhancement. While on the rubber wood dataset with more complex defect problem, the best mAP results were 94.7% with a speed of 139 FPS using the YOLOv5s-Transformer model with the addition of image enhancement Real ESRGAN.Keywords: detection, wood, Convolutional Neural Network (CNN), YOLO
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.