Vol. 38, Issue 4, pp. 737-754 (2008)

Abstract

A method for detecting forest fires smoke using SeaWiFS (Sea-viewing Wide Field-of View Sensor) images is developed in this paper. The colour masking technique is proposed to extract the maximum fires smoke pixels from the SeaStar/SeaWiFS satellite images by using Fusion by Arithmetic Combination (FAC) of the spectral bands method. Each image used is converted from RGB (Red, Green, Blue) to HIS (Hue, Saturation, Intensity) system. The resulting smoke plumes pixels are obtained visually in the Intensity and Saturation images. Then the values of intensity and saturation are analyzed to be potentially applied in other images. In this research, we applied our detecting forest fires smoke algorithm in seven different scenes, and in a variety of conditions, including different regions of the planet, and different dates. Next, Smoke Pixel Reference Ratio (SPRR) was used to test the proposed method. We found that the method can detect maximum pixels of smoke plumes in spite of some limitations.