Research on Hail Recognition Algorithm of FY-2G Satellite Data Based on Gaussian Distribution and SVM Model
Peng Yuxiang
Liu Tao
Yang Jing
Zhou Yongshui
Li Hao
Tang Piru
Li Huaizhi
Wen Jifen
Abstract：Based on Gaussian distribution and SVM models,this article conducts research on hail identification indicators and algorithms based on seven inversion products from the FY-2G satellite.Using data from 368 sets of FY-2G satellite inversion products for both hail and non-hail spots in 30 hail days in Guizhou Province from 2020 to 2022,and Gaussian distribution,we obtain quantitative hail iden-tification indicators.L-SVM,RBF-SVM,and S-SVM models are established to conduct hail identifica-tion research.The results show that quantitative hail identification indicators can be established for six in-version products based on Gaussian distribution,namely,cloud top height,cloud top temperature,supercooled layer thickness,optical thickness,liquid water path and black body brightness temperature.The SVM models based on the three kernel functions can effectively identify hail for both hail and non-hail spots,with accuracy rates exceeding 70%.Among them,the RBF-SVM model has the highest accu-racy rates for the total sample and non-hail spot samples,at 87.50%and 91.85%,respectively.The S-SVM model performs best for hail spot identification(89.13%).
Keywords：Gaussian distributionSVMhail recognitionkernel functionprobability density
Publication Date：2025-11-30
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：9( 81-89 )