An Intelligent Computing Prediction Model for Satellite Images
Using Empirical Orthogonal Function (EOF) method, the time coefficients were extracted from the samples of infrared satellite images every 3-h in heavy rainfall processes as predictands for images prediction modeling. Based on the technique of the reduction of data dimensionality, genetic neural network ensemble prediction (GNNEP) models have been developed for the associated predictands using predictors from physical quantities prediction products of numerical prediction model. The future satellite images were obtained by integrating the predicted time coefficients with the corresponding space vectors. Results show that the nonlinear prediction model can better forecast the main features of the development of cloud cluster with heavy rainfall in future 20-h.
sateWite image intelligent computing genetic algorithm heavy rainfall
Long Jin Ying Huang Ru He
Guangxi Climate Center Guangxi Meteorological Bureau Nanning, 530022, China
国际会议
昆明、丽江
英文
1314-1318
2011-04-15（万方平台首次上网日期，不代表论文的发表时间）