A new algorithm for forest fire smoke detection based on MODIS data in Heilongjiang province

Smoke is emitted from biomass burning every year. It affects regional air quality and long-term climate and is an important factor for fire detection of remote Sensing. An improved smoke plume detection method based on Kmeans clustering and multi-spectral threshold analysis is described in this paper. The threshold method has been developed with moderate-resolution imaging spectroradiometer (MODIS) spectral bands based on the analysis of spectral characteristics of different cover types. The image is categorized into two major types initially by Kmeans method on the basis of landmark spectrum analysis. The first class includes clouds, smoke and so on, and the second class includes vegetation, water and land. Then a multi-spectral threshold detection is applied to eliminate interference such as cloud for the first class and the automatic smoke identification is provided at last. The results have been validated with false color images for a number of cases from different areas and time, showing that the algorithm works well except for a few missing or incorrect identified smoke pixels.