Deteksi Api Berbasis Sensor Visual Menggunakan Metode Support Vector Machines

Muzakkiy, Hamdi Ahmadi (2016) Deteksi Api Berbasis Sensor Visual Menggunakan Metode Support Vector Machines. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
5112100091-Undergraduate Thesis.pdf

Download (2MB) | Preview

Abstract

Kebakaran adalah salah satu bencana yang sering terjadi. Penyebab sering terjadinya kebakaran yaitu karena kelalaian manusia, dan hubungan arus pendek listrik. Bencana kebakaran tidak hanya merusak bangunan bahkan menimbulkan banyak korban. Saat ini banyak alat pendeteksi api menggunakan sensor panas, ion, infrared. Namun penggunaan sistem alarm ini tidak akan bekerja hingga partikel mencapai sensor. Oleh karena itu diperlukan sistem deteksi api yang dapat mendeteksi kebakaran dengan cepat. Dalam Tugas Akhir ini diimplementasikan perangkat lunak pendeteksi api menggunakan deteksi gerak, deteksi warna menggunakan probabilitas warna, region growing, ekstraksi fitur wavelet dan klasifikasi piksel menggunakan support vector machines. Hasil dari deteksi bentuk akan digunakan dalam proses penentuan api. Dataset yang digunakan dalam proses uji coba berisi enam puluh tujuh video dengan panjang video enam sampai enambelas detik yang diambil dari berbagai sumber. Performa terbaik yang dihasilkan adalah true positif sebesar 96.32%, false positif sebesar 1.46% dan missing rate sebesar 2.23% =============================================================================================== Fire is one of the disasters that often occur. The cause of frequent occurrence of fires are due to human negligence and short-circuits. Fire does not only damaged buildings even cause many victim. Currently many of fire detectors use heat sensors, ion, and infrared. However, the use of this alarm system will not work until the particles reach the sensor. Therefore, there has to be a fire detection system that can detect fires quickly. In this final project fire-detection software is implemented using motion detection, color detection using color probabilities, region growing, features extraction using wavelet and pixel classification using support vector machines. The Results from shape detection will be used in the process of determining the fire. The dataset used in the testing process contains sixtyseven videos with a length of six to sixteen second video taken from various sources. The resulting performance is the best at 96.32% true positive, 1.46% false positif and missing rate of 2.23%.

Item Type: Thesis (Undergraduate)
Additional Information: RSIf 621.367 Muz d
Uncontrolled Keywords: Deteksi Gerak, Deteksi Warna, Probabilitas, Wavelet, Support Vector Machines
Subjects: T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques
Divisions: Faculty of Information Technology > Informatics Engineering > (S1) Undergraduate Theses
Depositing User: Mr. Tondo Indra Nyata
Date Deposited: 04 Oct 2017 08:41
Last Modified: 27 Dec 2018 03:03
URI: http://repository.its.ac.id/id/eprint/48918

Actions (login required)

View Item View Item