基于时序遥感卫星云图的对流云团动态追踪预测

Identifying and Tracking Convective Clouds Based on Time Series Remote Sensing Satellite Images

利用气象遥感红外云图识别、追踪和预测对流云团的发展变化将有助于及时预报灾害性天气,但目前仍然缺乏有效的动态追踪预测方法.本文采用多帧时序遥感云图,提出了一种对流云团识别、跟踪和趋势预测的实时快速预测模型.通过模糊C均值(FCM)算法识别对流云团,然后对输出结果使用优化的近邻交叉相关法动态追踪,并根据云团运动的气旋学说,创新性的采用三次样条插值函数拟合云团运动路径,实现云团的动态追踪预测.实验结果表明该方法达到了较高的预测准确率.

Identifying and tracking convective clouds from remote sensing images is very useful for timely forecasting the se -vere weather .However ,very limited methods have been proposed for this purpose .This paper proposes a method that can automati-cally identify ,track the convective clouds and then predict the movement trends .First ,the fuzzy C-means （FCM） algorithm is used to identify and segment the cumulonimbus from the cloud images ;then ,the near neighbor cross-correlation method is used to track the convective clouds ;finally ,in accordance with the movement characteristics of the aerodynamic flow ,cubic spline interpolation functions are used to fit the clouds’ movement paths ,by which the movement trends of the convective clouds can be predicted .The contrast results to the real images show that the proposed method can achieve satisfying prediction accuracy .