本文应用BP神经网络算法对铂电阻温度传感器进行非线性校正，给出了BP神经网络的结构和训练权值的方法，在训练网络时对输入量进行了归一化处理。并应用此训练的网络对一实际的温度采集系统进行校正。此方法实现简单，大大方便了铂电阻温度传感器在温度测量中的应用。

This article uses BP neural network to calibrate nonlinear characteristic of platinum resistance, gives the structure of BP neural network and the method to train weight. The input are normalized when train the network. Finally, this neural network calibrates a fact system. This method can be achieved very easily. Using platinum becomes very convenience in measurement through this method