FY-3D infrared hyperspectral cloud detection method based on extreme random tree

Abstract

The invention discloses an FY-3D infrared hyperspectral cloud detection method based on an extreme random tree, and relates to the technical field of satellite remote sensing. According to the method,HIRAS and MERSI carried on FY-3D are used for continuously observing data of the earth, time matching is conducted on two data files, space matching is conducted on pixels of two instruments, cloud labels of HIRAS pixels are determined through the cloud labels of the matched MERSI pixels, HIRAS data with the cloud labels are obtained, and a training data set of an extreme random tree model is formed; a data set is trained by using an extreme random tree algorithm, the performance of the model is checked by using a test data set, and finally the extreme random tree model is applied with bettercloud detection performance obtained by training to cloud detection processing of YF-3D infrared hyperspectral data. The cloud detection method is short in cloud detection time, low in cost and highin cloud detection classification accuracy.

Description

technical field [0001] The present invention relates to the field of satellite remote sensing technology, in particular to a decision tree-based integrated learning method in the field of machine learning—extreme random tree algorithm, combined with the method of detecting clouds in the field of view of HIRAS by using infrared hyperspectral data of FY-3D satellite HIRAS instrument . Background technique [0002] Infrared hyperspectral data is an important observational data for modern numerical weather prediction systems. Assimilating infrared hyperspectral data into a near-model system can effectively improve the forecasting level. Water droplets and ice crystals in clouds absorb infrared radiation so effectively that satellite infrared detectors cannot detect infrared radiation from the atmosphere below the clouds and the surface. In addition, it is difficult for current radiative transfer observation operators to accurately simulate the radiative effect of clouds. There...

AI Technical Summary

Problems solved by technology

[0003] At present, the multi-channel threshold method proposed based on the physical characteristics of clouds is applied to most commercial satellite infrared sounders, but many cloud detections have multiple thresholds, and it is difficult to set an appropriate threshold; there is another widely used method is Imager-based cloud detection method, although this method significantly improves the cloud detection accuracy of hyperspectral infrared detection, but this method requires high memory cost and computing time cost

Images