面向遥感卫星数据获取应用的时间序列云量预测方法  

A Cloudiness Prediction Method for Remote Sensing Satellite Data Acquisition Applications

作　　者：王志信[1,2] 黄鹏[2] 林友明[2] 贾秀鹏[2] 

机构地区：[1]中国科学院大学,北京100049 [2]中国科学院遥感与数字地球研究所,北京100094

出　　处：《遥感信息》2014年第3期8-13,共6页Remote Sensing Information

基　　金：863计划(Y28003102A);国家科技支撑计划课题(Y26Z02101A-1)

摘　　要：针对目前没有能够直接满足遥感数据获取需求相关的云量预测方法的问题,本文提出了一种利用时间序列分析预测方法对云量进行预测的方法。在云量特征分类的基础上,整合ARMA、ARIMA和SARIMA 3种模型对云量进行预测,并得到了满意的预测结果。根据云量预测结果参考信息,选择云量覆盖较少的时间段进行卫星成像规划,有利于更合理的进行卫星成像规划,对满足遥感数据获取需求有重要意义。There is not any cloudiness prediction method that can directly meet the needs of remote sensing data acquisition so far.In order to solve this problem,a method using time series analysis and forecasting was proposed to predict cloudiness.In this work,we integrated three models,which are so called ARMA,ARIMA and SARIMA,for predicting cloudiness based on cloudiness feature classification,and the results show that the accuracy of prediction is satisfactory.With the information of cloudiness prediction,it would be easier to do satellite imaging planning by selecting the time period when cloudiness coverage is less and it would also help to meet the demands of remote sensing data acquisition.

关 键 词：云量预测 时间序列 卫星成像规划 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]