Aiming at measurement errors of tipping bucket rain gauge sensor,an effective measures is proposed and three types of curve fitting algorithms are evaluated in predicting rain gauge data of Beijing 7. 21 rainstorm event observations,by the comparison of their predicting performance. The experimental results show that the prediction effect of general regression neural network (GRNN) algorithm is the best of the three. Since there are some limits to use rain gauge data by meteorological department,for which an intelligent information processing scheme is also proposed based on the predicted data,which can entirely make the rainfall information more accurate,intuitive and comprehensive,improving the coping abilities of meteorological department for emergency events.%针对当前翻斗式雨量计传感器存在的计量误差,提出了有效的应对措施,并使用3类曲线拟合算法对北京市7.21暴雨实际观测资料的雨量计数据进行预测和比对.实验结果表明:广义回归神经网络(GRNN)算法预测效果最好.针对气象部门对雨量计测量数据的使用方式局限性,据此结合预测数据给出了一种智能信息处理的方案,能够将雨量信息较为准确、直观和全面地展现出来,增强了气象部门对应急事件的处理能力.
展开▼
