%0 Journal Article
%@ 18196608
%A Kadir, E.A.
%A Rosa, S.L.
%A Othman, M.
%A Daud, H.
%D 2021
%F scholars:14601
%I Asian Research Publishing Network
%J ARPN Journal of Engineering and Applied Sciences
%N 15
%P 1616-1622
%T Smart sensor system to classify hotspot types potentially for land and forest fires
%U https://khub.utp.edu.my/scholars/14601/
%V 16
%X A fire hotspot exhibits the potential to create forest and wildfire, and the size of a hotspot determines the potential level to become a fire and its spread rate. Wild and forest fire is a major issue in some counties with a large forest area, especially in a tropical country, such as Indonesia. This research aims to identify and classify the fire hotspot types and their potential to become a large fire that spreads to forest and wild in a tropical region. A sensor detection system is developed to detect the type of fire hotspots. Several sensors are used to identify and classify the model and type of hotspots and their potential level to become a fire that threatens the wild and forest. The fire sensor is used as the main sensor to detect a fire, and other sensors are utilized to obtain supporting data, such as temperature, humidity, and carbon. A computer algorithm is used to classify the types of hotspot potential to spread to the forest on the basis of the data received from all the sensors, especially the fire sensor. The data received from the carbon sensor are used as parameters to determine whether a hotspot can cause a fire or not. Results show that the proposed sensor system can differentiate and classify whether the hotspot has potential to become a fire or only a small and controllable hotspot. The system can also classify the hotspot data in actual condition, including noises, such as flashlight, touch light, and hotspot from cigarette matches. The decision from the sensor system is extremely effective in assisting for forest fire preventive action rather than conventionally shutting down the fire in every hotspot detected. © 2006-2021 Asian Research Publishing Network (ARPN). All rights reserved.
%Z cited By 0