Spatial analysis of thermal anomalies from airborne multi-spectral data

Abstract

Based on field thermal measurements, thermal anomalies caused by coal fires can be grouped into three categories, low- (up to 20°C above the background), medium- (20–120°C) and the high-amplitude (above 120 °C) ones. Night-time airborne thermal scanner data acquired in the 8–12.5 µm wavelength region of the electromagnetic spectrum clearly show the coal fire areas and the background areas. However, one of the disadvantages of the night-time data is the saturation of the medium- and high-amplitude thermal anomalies. In the daytime 8–12.5 µm image, the medium-amplitude thermal anomalies can be detected. These thermal anomalies only partly represent underground coal fires and some of them represent the solar-heated coal seams and black shale with higher emissivity. Daytime thermal infrared data acquired in the 3–5 µm wavelength region provide information from both the spectrally reflected solar radiation and radiation from high-intensity surface thermal anomalies of the underground coal fires. To reduce the effects of the spectrally reflected solar radiation, the data acquired in the 0.61–0.69 µm wavelength region were used to adjust the 3–5 µm data and the new image shows the enhanced high-amplitude thermal anomalies of the underground coal fires. The three kinds of data have been fused to integrate the background, low-, medium- and high-amplitude thermal anomalies, which are highly correlated to the field thermal measurements. On the basis of the spatial patterns of the thermal anomalies and the underground coal fire spreading models set up through field observation, the spreading direction of underground coal fires is inferred. Comparing the daytime and night-time 8–12.5 µm data, the solar-heated coal seams were detected as areas of high-risk for coal fires occurring in the future, because the temperatures in these areas measured in the field were close to the critical point of the spontaneous combustion of coal.

Acknowledgments

This research was carried out under a European Commission Research Contract No. CI1*-CT93-0008 (DG HSMU). The authors are grateful for financial support from the ITC Research Programme (project No. P3063315). Mr Wan Yuqing carried out the image processing of figure 5(a). The authors also gratefully acknowledge the assistance of Dr Tan Yongjie, Prof. Zhang Tingan and Mr Kang Gaofeng in arranging the field work during 1994 and 1995.