电容式降雨传感器及其特性曲线拟合方法

摘要

Aiming at problem of large nonlinear error of characteristic curve of capacitive sensor applied to measurement of rainfall,least square curve fitting method and radial basis function(RBF) neural network are used to fit output characteristic curve. The results show that the RBF neural network model has stronger nonlinear mapping ability,and its fitting precision is significantly higher than that of the least square polynomial model.%针对将电容式传感器应用于降雨测量时,特性曲线的非线性误差较大问题,分别采用最小二乘曲线拟合法和径向基函数(RBF)神经网络对其输出特性曲线进行拟合.结果表明:RBF神经网络模型具有更强的非线性映射能力,其拟合精度明显高于最小二乘多项式模型.