基于高密度示踪云选取的云迹风反演技术研究

Deriving Cloud Motion Winds from Geostationary Meteorological Satellite Images Based on Selecting High Density Tracers

基于相关性判别因子(即示踪云追踪或图像匹配方法)的理论分析研究,以及3种追踪方法运用于云迹风反演的对比研究,探索出运用梯度分析方法进行示踪云选取的一种可能依据,并给出梯度的一种离散计算公式,另外还对该示踪云选取方法进行了补充和完善。对完善后的该方法选取的示踪云进行云迹风反演,并对其进行风场分析,并与无线电探空风场进行比较。结果表明,云迹风质量好、密度高,能够清晰地显示出天气系统的风场结构。

Based on the research of theoretical analysis on three pattern recognition parameters（i.e.tracking or pattern matching techniques）,and the work of contrasting three tracking techniques applied in deriving cloud motion winds,apossible warranty of the method of gradient analysis tracer selection can be got.A discrete formula for computing gradient is also provided.Further more,the method of gradient analysis tracer selection is consummated.Cloud motion winds are derived from the tracers selected by the consum-mated method and after a wind field analysis on the cloud motion winds,the cloud motion winds are com-pared with the rawinsonde winds.The result shows that the cloud motion winds have a good quality and a high density,and the wind structure of weather system can be seen clearly from them.