A Study On The Data Processing And Method For Thermal Infrared Anomalies Associated With Earthquakes Based On MODIS LST Products

Earthquake’s severely destruction and unpredictability making it one of the most terrible natural disasters human has faced for a long time, seismic activity is in active in China, earthquake cause serious socio-economic losses and people’s lives and property losses. Studying earthquakes and reducing earthquake disaster is an important task for China. The traditional means of monitoring earthquakes appears constrained by the large scale of geological seismic activity, remote sensing and other new technologies enhance human’s capabilities in macroscopic observation of the earth surface and the mass obtaining of information. Earthquake thermal infrared anomalies were found in the late 1980 s, researchers from many countries raised concerns, and consequently thermal infrared remote sensing found bonding point with the seismic observation.In this paper, based on previous studies, we made a test using RST(Robust Satellite Techniques) and its improved algorithms to process the MODIS LST(Land Surface Temperature) data, extracting earthquake thermal anomalies, and compare the results of different characteristics of each algorithm, accumulate experience, then enhance the continuous improvement of the thermal anomaly detection results. Due the incompleteness of data caused by the impact of cloud, time series reconstruction method is used to complement the missing data. Apply the extraction algorithms mentioned above to the test data before and after reconstruction, then compare the results and examine the effect of data reconstruction.The experimental results showed that:(1) two RST based on time domain average – RST based on average over the same time each year and RST based on year average, have a similar results, so use any one as same.(2) RST based on time domain average can overcome the lack of sensitivity of abnormal warming of a large area with RST based on spatial domain average is, but is too sensitive to interannual difference, resulting in suppression and interference earthquake thermal anomalies. Two algorithms can be cross-referenced to help improve the accuracy of detecting thermal anomaly.(3) LST time series reconstruction algorithm used in this article- Savitzky-Golay filtering algorithm can make the result data in on time and space have better characteristics, partly to enhance the effectiveness of the thermal anomaly extraction algorithm, but data reconstruction error cause errors and noise in the extraction algorithm results.(4) RST based on year trends partly make up for the shortcomings of the previous algorithms, is neither too sensitive to interannual difference, nor inhibit a large area of high value ALICE warming, but it still needs more fine-grained data reconstruction method to reduce noise and improve its effect.(5) The distribution characters of Earthquake thermal infrared anomalies is closely associated with the faults distribution and the level of faults activity. The precise characteristics of these relationship which can be very valuable is expected to be available through a lot of earthquakes information and more efficient algorithm theory.