小波熵在植物电信号识别中的应用研究

摘要

By applying the better time-frequency localization ability of wavelet and the ability of entropy able to token system state, the wavelet entropy is taken as the eigenvector of plant electrical signals. Then the feature is put into BP neural network classifier for autorecognition, with good recognition effects. Obtained the wavelet energy entropy is used to analyze the results. These results also show that the wavelet entropy eigenvector is more efficient than the energy eigenvector of wavelet coefficient in the plant electrical signal recognition.%利用小波良好的时频局部化特性以及熵能够对系统状态进行表征的特点,提出将小波熵作为植物电信号的特征向量,将该特征输入到BP神经网络分类器进行自动识别,取得了良好的识别效果.同时,利用小波能量熵对结果进行分析.结果表明,小波熵比小波系数能量作为特征对植物电信号的识别更有效.