Algorithm for statistical downscaling of land surface temperature using ElasticNet

Nowadays there are many algorithms for statistical downscaling of land surface temperature (LST) that can improve the spatial resolution of this data (Zhan et al., 2013). The problem of existing algorithms is the necessity to select features and adjust the algorithm parameters depending on land surface classes present in the image. The objective was to develop an algorithm for downscaling land surface temperature with parameters minimally dependent on the land surface classes.