奇异值分解与神经网络结合的卫星云图云团移动预测  

Cloud cluster movement forecast technique of satellite cloud pictures based on singular value decomposition and artificial neural networks

作　　者：刘科峰[1] 张韧[1] 李文才 赵中军 江海英 

机构地区：[1]解放军理工大学气象学院 [2]解放军92493部队

出　　处：《解放军理工大学学报（自然科学版）》2008年第3期298-301,共4页Journal of PLA University of Science and Technology(Natural Science Edition)

基　　金：中国博士后科学基金资助项目(2004036012);江苏省博士后科研基金资助项目(0401068B)

摘　　要：云团运动和发展演变的预测是暴雨等灾害性天气监测预报的重点和难点问题,针对当前云团预测中缺乏有效的非线性、非平稳预测手段,提出了奇异值分解SVD(singular value decomposition)与径向基网络相结合的云团预测途径。首先用SVD对云图进行分解,提取主要的云团结构特征,然后用提取出的云图奇异特征值和左右奇异向量作为模式识别因子,选择特定区域和季节的云图时滞序列采样,并用前后时段样本云图的奇异值和奇异矢量作为云图预测模型的输入、输出,通过对径向基网络的学习训练和误差迭代收敛,建立了云团运动的非线性预测模型。试验结果表明,该方法能合理地描述云团运动的基本特征和演变趋势。The forecast of cloud cluster movement and evolution is important ,which is difficult for inspecting and predicting such disaster weather as rainstorm. In view of lacking effective non-linear and non-stable cloud cluster movement forecast technique, a cloud cluster forecast new technique based on singular value decomposition （Singular Value Decomposition, SVD） and radial basis neural networks was presented in this paper. First, the satellite cloud picture was decomposed by SVD, and its chief characters were pickedup, then the picked singular characteristic value and its right/left singular characteristic vectors were taken as pattern identification factors, the satellite cloud picture time series were sampled in specified region and season, and the fore-/back-period samples singular values and singular vectors were taken as the input/output of the forecast model. By training the ANN model, a non-linear forecast model of cloud cluster movement was established. The results show that the modeling technique can reasonably describe the basic characters and the evolvement trend of cloud cluster.

关 键 词：云团预测 奇异值分解 径向基网络 

分 类 号：P412.27[天文地球—大气科学及气象学]