Application of forest fire detection index in forest fires: a case study of Aberdare forest
View/ Open
Date
2018Author
Aseko, Japheth Omondi
Type
ThesisLanguage
enMetadata
Show full item recordAbstract
Forest fire detection is an important step in wildfire pre-suppression process. Many organizations
involved in forest monitoring and surveillance strive to achieve efficient and faster methods to
detect forest fires. This is due to colossal amount of money spent in forest restoration and
rehabilitation after fire occurrences. Timely detection and efficient interpretation of the satellite
and other terrestrial images allow fire personnel to contain the fire at its initial stage and this will
reduce damages caused to both flora and fauna besides the suppression cost. It has been difficult
to predict, detect and control forest fires in Kenya and in many cases different agencies are
caught unawares in the instances of fires witnessed in the country. Forest fires spread faster due
to the presence of dry vegetation covers especially during dry season and this requires a faster
way of detection to avoid escalated damages as witnessed in the previous fire occurrences in our
country; the latest being the one in Aberdare forest which formed the basis for this project. The
motivation behind this project was to have a smooth means for forest fire detection through the
use of a forest fire index that will help in detecting flames and smokes in the near-real time and
to enable concerned parties to react quickly in case of forest fire. The method borrowed from the
vegetation classification to detect distinct reflectance from both flame and smokes. Smoke has
been difficult to detect because of the weak infrared radiations emitted at its production and this
makes sensors not to detect it. In this project smokes was detected adaptively in the region of
interest with the help of a variable factor. The project was carried out in phases; MODIS terra
and aqua image acquisition from https://ladsweb.modaps.eosdis.nasa.gov/search/ and terrestrial
image from drone on 30th March 2018, extraction of colour components, normalization of the
colour components, using the forest fire detection index for smoke and flame detection and lastly
and representing the output from colour analysis for smokes and flames in image forms. The
method can be used in near-real time forest fire detection resulting in a more cost-effective
outcome than the traditional systems involving water tanks and watch towers. This method is
efficient in forest fire detection at the early stage when both satellite and drone images are made
readily available.
Citation
Master of Science in Geographic Information SystemPublisher
University of Nairobi
Rights
Attribution-NonCommercial-NoDerivs 3.0 United StatesUsage Rights
http://creativecommons.org/licenses/by-nc-nd/3.0/us/Collections
The following license files are associated with this item: