Multi-spectral remote sensing based land surface temperature retrieval and isolated permafrost zone segmentation

Affected by global climate change, the permafrost in northeast high latitude permafrost area of our country was degraded significantly, and the water phase change seriously affected the mechanical indexes of soil during the thawing process of frozen soil. Due to the permafrost melting of high latitude permafrost regions in China in recent years, a plurality of freeway roadbed has landslide, and uneven subsidence and other geological disasters had repeated in the northern part of Heilongjiang Province. In order to study the island permafrost distribution, choosing the K153-K183 Road area from Bei'an to Heihe expressway as the study area, the 3 Landsat 7 ETM+ image data converging the study area was choosen respectively on March, May, September in 2009. The surface temperature map of the corresponding date was obtained and the relationship between ground temperature and NDVI, DEM was analyzed through the retrieval of the infrared surface temperature in the study area using the radiation transfer equation method. Then the map layer of the inversion results was classified using natural breakpoint classification. The distribution map of the low temperature area of the corresponding moment was obtained through the analysis and the artificial interpretation of the surface temperature change of the study area. Frozen soil distribution map was obtained in the study area in 2009 on the intersection operation of the low temperature zone. It had a good consistency comparing with the expressway permafrost investigation results from Bei'an to Heihe.