Method and device for generating all-weather surface temperature based on machine learning

Abstract

The invention discloses a method and device for generating all-weather surface temperature based on machine learning. The method uses MODIS tool MRT to extract the MODIS data set retrieved by remote sensing; and uses static meteorological satellite data and DEM terrain data of ALOS satellite Combined to estimate and obtain the incident solar radiation on the surface; spatially aggregate data sets of the same spatial scale, and use the above MODIS data set as a machine learning training data set; construct a surface temperature relationship model through a random forest model; estimate cloud coverage images The real surface temperature of the pixel; the real surface temperature of the cloud-covered pixel is combined with the data set of the cloud-free pixel to generate the all-weather surface temperature. The method of the invention solves the problem that the current thermal infrared remote sensing is easily affected by clouds and fog, and there are a large number of blank and missing areas in surface temperature products, realizes the estimation of surface temperature under cloudy conditions, and provides an important basis for the generation of all-weather surface temperature products.

Description

technical field [0001] The invention relates to surface temperature monitoring technology, in particular to an all-weather surface temperature generation method and device based on machine learning. Background technique [0002] Land surface temperature (LST), as an important parameter reflecting the interaction between the earth and the atmosphere in the earth's surface system, is a key parameter affecting the processes of surface ecology, hydrology, meteorology, etc. comprehensive results. Therefore, quantitative and accurate acquisition of the temporal and spatial distribution characteristics of surface temperature has important research significance and value for the study of the energy balance of the earth-atmosphere system and ecosystems. Moreover, dynamic monitoring of resources and the environment on a regional and global scale requires comprehensive, complete, and continuous all-weather information on the temporal and spatial distribution of surface temperature, su...

AI Technical Summary

Problems solved by technology

[0004] However, due to the low accuracy of passive microwave surface temperature retrieval, the large scale difference between passive microwave data and thermal infrared data, the error of passive microwave surface temperature itself and the uncertainty of downscaling, there are many problems in the all-weather surface temperature product generated by data fusion. large uncertainty; moreover, the existing problem of using the existing inversion algorithm is that the estimated surface temperature value is not the real surface temperature under the cloud at the satellite transit time, but the theoretical value of the surface temperature under clear sky conditions at the satellite transit time , it is difficult to meet the needs of practical applications

Images