DETEKSI KEBAKARAN BERDASARKAN VIDEO MENGGUNAKAN MASK R-CNN

DINATA, MAHENDRA and Rini, Dian Palupi and Rachmatullah, Muhammad Naufal (2025) DETEKSI KEBAKARAN BERDASARKAN VIDEO MENGGUNAKAN MASK R-CNN. Undergraduate thesis, Sriwijaya University.

Preview	Image RAMA_55201_09021282126066_cover.jpg - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (509kB) | Preview
	Text RAMA_55201_09021282126066.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (7MB) | Request a copy
	Text RAMA_55201_09021282126066_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (5MB) | Request a copy
	Text RAMA_55201_09021282126066_0023027804_0001129204_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (1MB)
	Text RAMA_55201_09021282126066_0023027804_0001129204_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (835kB) | Request a copy
	Text RAMA_55201_09021282126066_0023027804_0001129204_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_55201_09021282126066_0023027804_0001129204_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_55201_09021282126066_0023027804_0001129204_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy
	Text RAMA_55201_09021282126066_0023027804_0001129204_06.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (94kB) | Request a copy
	Text RAMA_55201_09021282126066_0023027804_0001129204_07_ref.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (109kB) | Request a copy
	Text RAMA_55201_09021282126066_0023027804_0001129204_08_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (127kB) | Request a copy

Abstract

Fire is one of the disasters that can cause great losses, both in terms of loss of life and economic loss. Fast and accurate fire detection is the main key to prevent fires from getting bigger and out of control. This study aims to develop a fire detection system using a video camera that can provide fire information as quickly as possible. This fire detection system uses Mask R-CNN model to detect fire objects in each frame captured by the video camera. The Mask R-CNN model was trained using 53184 images containing fire objects and objects that resemble fire. This image data was obtained from the results of augmenting a dataset consisting of 2216 images. The augmentation performed is an image rotation of 15 degrees from 0 to 360 degrees, so that the data produced after augmentation is 24 times larger. The results of the study showed that the developed model was able to detect fires with an AP50 of 77,87%. This result was obtained from several model experiments that produced the best model with a ResNet 101 backbone, 1000 number of proposals, and a base learning rate of 5 x 10^-4.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Fire Detection, Augmentation, Mask R-CNN
Subjects:	T Technology > T Technology (General) > T1-995 Technology (General)
Divisions:	09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User:	Mahendra Dinata
Date Deposited:	19 May 2025 08:34
Last Modified:	19 May 2025 08:34
URI:	http://repository.unsri.ac.id/id/eprint/173265

Actions (login required)

View Item