MUSYAFFA, NUNO DZAKKII (2026) ANALISIS PERFORMA MODEL DEEP LEARNING BERBASIS CONVOLUTIONAL NEURAL NETWORK FIRENET DAN YOLOV8 DALAM COMPUTER VISION UNTUK DETEKSI KEBAKARAN. S1 thesis, Fakultas Teknik Universitas Sultan Ageng Tirtayasa.

	Text (Fulltext) Nuno Dzakkii Musyaffa_3332210031_Fulltext.pdf Download (3MB)
	Text (Bab 1) Nuno Dzakkii Musyaffa_3332210031_01.pdf Download (783kB)
	Text (Bab 2) Nuno Dzakkii Musyaffa_3332210031_02.pdf Download (571kB)
	Text (Bab 3) Nuno Dzakkii Musyaffa_3332210031_03.pdf Download (347kB)
	Text (Bab 4) Nuno Dzakkii Musyaffa_3332210031_04.pdf Download (1MB)
	Text (Bab 5) Nuno Dzakkii Musyaffa_3332210031_05.pdf Download (179kB)
	Text (Daftar Pustaka) Nuno Dzakkii Musyaffa_3332210031_Ref.pdf Download (195kB)
	Text (Lampiran) Nuno Dzakkii Musyaffa_3332210031_Lamp.pdf Download (1MB)
	Text (Dokumen Hasil Cek Plagiasi) Nuno Dzakkii Musyaffa_3332210031_CP.pdf Download (22MB)

Abstract

Fire incidents are among the most frequent disasters occurring in both urban and rural areas. With the advancement of Artificial Intelligence and Computer Vision technologies, new methods for fire detection through imagery have emerged by leveraging the capabilities of Convolutional Neural Networks (CNN). CNN-based models such as FireNet and YOLOv8 are promising approaches for identifying the presence of fire within an image. In this study, both models, after undergoing a training process, were evaluated using five video footages with varying characteristics. The accuracy values for the FireNet model on Footage 1, Footage 2, Footage 3, Footage 4, and Footage 5 are 89.67%, 98.78%, 91.22%, 98.22%, and 63.67%, respectively. Meanwhile, the accuracy values for the YOLOv8 model on Footage 1, Footage 2, Footage 3, Footage 4, and Footage 5 are 98%, 69.44%, 81.22%, 85%, and 99.67%, respectively.

Actions (login required)

View Item