Extracting thermal anomalies of underground coal fire from multi-temporal daytime images

Coal mine spontaneous burning has led to resources waste, meanwhile it also causes air pollution and some geological disasters. Now remote sensing is becoming an important approach of monitoring the coal fire. The thermal anomalies points are extracted by three methods in this article that is multi-temporal MNF transformation, multi-threshold density slice and differentiated method. Combining with GIS the information of the coal fire extracted are qualified compared and analyzed, then find the best method to extract the high thermal anomalies points is the difference method, and helped by the MNF transformation areas where affected by coal fires extracted are more accuracy than directly using density slice.