基于图像挖掘技术的降雨预测系统的研究

目前，计算机图像识别技术已经被广泛地应用于天气和降雨预测等气象预报部门，但是对于短期和局部的降雨预测而言，其准确性和快速性都无法满足人们的需求；基于此，提出了一种基于卫星云图挖掘的降雨预测应用系统，通过提取雨量站的局部图像特征以及相关联的降雨量，获得其强关联规则以此作为降雨预测的先验知识库；最后通过对当前云图的特征提取，并依照先验知识库进行推理，最终给出预测结果；实验结果表明，该系统有比较好的实用效果。

Precipitation Predicting System Based on Satellite Image Mining

Present automatic precipitation predicting systems are widely adopted in the weather and precipitation predicting, but their accuracy and efficiency are not adequate for local and short time precipitation predicting. So, a system based on satellite image mining technology is presented. After obtaining the local characteristics of precipitation sampling point and the actual precipitation, the system mines these data to set up a knowledge base, which is used for later predicting. When current image received, its characteristics are retrieved which is in turn used to for referring. Our experiment shows that the method has been satisfying practical results.