An Improved Algorithm for Forest Fire Detection Using HJ Data

摘要

Using the characteristics of the environment and disaster monitoring and forecasting small satellite constellation (HJ), the moderate resolution imaging spectroradiometer (MODIS) forest fire detection contextual algorithm was improved to adapt the HJ-infrared sensor (HJ-IRS). The enhanced method consisted of potential fire pixel identification, absolute and relative fire pixel judgment, background characteristics analysis, and fire pixel confidence. The improved algorithm was programmed in IDL7.1 and tested using HJ forest fire data from Heilongjiang Province in 2009. Results show that improving the forest fire detection contextual algorithm to adapt HJ-IRS is feasible and highly accurate. HJ data are much more sensitive to smaller and cooler fires than MODIS or the advanced very high resolution radiometer (AVHRR) data, and the improved capabilities offers a good potential for application in forest fire detection.