【作者】 赵在宽；

【导师】 舒辉；

【作者基本信息】 解放军信息工程大学 ， 计算机应用技术， 2009， 硕士

【摘要】 河南作为一个农业大省,干旱对经济,特别是农业生产的影响很大。干旱程度是对农作物生长形势研判的一个重要指标。在河南农业气象研究中,主要利用土壤相对湿度干旱指标来判定干旱程度。它是在测量土壤湿度后,计算相对湿度所得。目前,河南省气象部门测量土壤湿度是采用人工取土烘干称重的方法,工作量大,频次少,不利于对农作物生长形势的判定及农气研究。多年来,河南省气象部门积累了大量的气象观测数据。如何综合利用这些信息,进行土壤相对湿度判定,为河南农业发展服务,是气象服务工作中的一个急需解决的问题。本文对此进行了比较详细的研究与实践,主要研究内容和创新点如下:1.分析了国内外数据仓库、数据挖掘技术现状和河南气象档案业务发展情况,根据河南实际情况和本文干旱指标挖掘研究需要,设计并初步建设了河南气象数据仓库。2.研究土壤相对湿度干旱指标的估测方法,首次利用BP神经网络和RBF神经网络分别对该指标进行了挖掘,形成了土壤相对湿度干旱指标挖掘模型,并利用模型对河南干旱情况进行了评估。更多还原

【Abstract】 Drought has an affect on the economy, especially great on agriculture in Henan, which is abig agricultural province. Drought grade is a key index to judge the crop growth. Soil RelativeHumidity Drought Index is mainly used to evaluate the soil drought grade in agrometeorologicalresearch in Henan. It is the result of soil absolute humidity after calculating soil humility. Atpresent in Henan meteorological department, the soil drought grade is evaluated by digging,drying and weighting soil, and the method is burdensome and of low frequency, so it is notenough for the agriculture research and crop growth estimation.For years, plenty of meteorological data are accumulated in Henan meteorologicaldepartment. Finding a way of estimating the soil relative humidity drought index with the data toserve the agriculture development of Henan is the goal of this thesis.The main works and creations are as the followings:1. By analyzing the status of data warehouse and mining technique and meteorologicalrecord development, Henan meteorological data warehouse is designed and preliminaryconstructed according to the situation in Henan and the research of drought index mining in thepaper.2. BP neural network and RBF neural network is firstly used to research soil relativehumility drought index and a soil relative humility drought index mining model is formed andused in estimating drought grade in Henan.更多还原