Universiti Teknologi Malaysia Institutional Repository

Early detection of potential forest fires using satellite remote sensing techniques

Mohd. Hassan, Aida Hayati (2008) Early detection of potential forest fires using satellite remote sensing techniques. Masters thesis, Universiti Teknologi Malaysia, Faculty of Geoinformation Science and Engineering.

[img] PDF - Submitted Version
Restricted to Repository staff only

2703Kb
[img] PDF (Abstract)
94Kb
[img] PDF (Table Of Content)
123Kb
[img] PDF (1st Chapter)
102Kb

Abstract

In 1997/1998, Malaysia experienced one of the most severe forest fire episodes in history as a consequence of a prolonged dry season following the El- Nino phenomenon. Since then, uncontrolled fires, atmospheric pollutions and haze remained as a common problem throughout the dry period in this region. The estimated cost of the damage caused by forest fires in Malaysia is about RM816.47 million a year. The loss by forest fire episodes has brought to light the importance of developing better tools for effective forest fire management systems. In this research, three sets of computer programmes were designed for: detecting hot spots; computing the fire risk index and generating spatial analysis for detected fires. Remote sensing and GIS techniques have both been integrated in this work. Eventually, a simple yet robust early warning system for forest fire detection in Malaysia has been devised. Thermal bands of MODIS (Moderate Resolution Imaging Spectroradiometer) were used to extract hot spot information and to generate a fire risk map. Proximity analysis was carried out using an extension in ArcView GIS software. The results from this research were compared with forest fire occurrence information from the Fire and Rescue Department of Malaysia (FRDM) and information of rainfall and temperature from the Malaysian Meteorological Services (MMS). High correlation (R2 = 0.8) was found between temperature derived from MODIS and the temperature obtained from the MMS. Forest fire map generated from the study also gave a high accuracy (71%). Normalized Difference Vegetation Index (NDVI) values derived from MODIS were found to be highly correlated (R2 = 0.7 and R2 = 0.85) with rainfall and temperature data obtained from the MMS, respectively. Hence, the output of the research shows that this system can be used as an early warning system mechanism to mitigate forest fire incidence and can be upgraded into a more complex system to enhance its functioning.

Item Type:Thesis (Masters)
Additional Information:Thesis (Sarjana Sains (Remote Sensing)) - Universiti Teknologi Malaysia, 2008; Supervisor : Prof. Dr. Mazlan bin Hashim
Uncontrolled Keywords:artificial satellites, remote sensing, fire detectors, environmental monitoring
Subjects:T Technology > TD Environmental technology. Sanitary engineering
Divisions:Geoinformation Science And Engineering (Formerly known)
ID Code:11443
Deposited By: Ms Zalinda Shuratman
Deposited On:15 Dec 2010 08:33
Last Modified:27 Dec 2013 03:52

Repository Staff Only: item control page