Active–Passive Remote Sensing Identification of Underground Coal Fire Zones With Joint Constraints of Temperature and Surface Deformation Time Series

Abstract

Coal fire is a geological disaster that causes resource waste and environmental pollution globally. Accurate identification of the spatial location of coal fires is critical for effective coal fire governance. However, existing methods for identifying coal fire zones have problems, such as a high omission and misclassification ratio and insufficient consideration of the temporal variation in temperature. Therefore, this article proposed a temporal temperature anomaly extraction algorithm based on adaptive windows (TTAE-AW) to extract temporal temperature anomaly information. Moreover, the spatial coverage of deformation monitoring points was improved using distributed scatterer interferometric synthetic aperture radar (DS-InSAR), and then a double-threshold two-stage filter method (DTTF) was proposed to accurately identify the spatial location of coal fire zones. The Rujigou mining area in Ningxia (China) was chosen as the region of study. Results showed that the temperature anomalies extracted using the TTAE-AW method are more concentrated in coal fire zones and that the amount in different seasons is more stable. The average accuracy and Kappa coefficient were improved by 15.5% and 0.345, respectively, over those of the conventional method. Compared with the small baselines subset InSAR approach, the DS-InSAR technique has 158% higher spatial coverage for monitoring coal fire zones. Compared with in situ observations of coal fire points, the accuracy and Kappa coefficient of the spatial location of fire zones obtained using the DTTF method were 91% and 0.77, respectively, demonstrating that the proposed method can provide reliable technical support for coal fire monitoring and management.

Keywords

• Collaborative analysis
• distributed scatterer interferometric synthetic aperture radar (DS-InSAR)
• surface deformation
• thermal infrared (TIR) remote sensing
• time-series temperature anomaly extraction (TAE)
• underground coal fire identification