摘要
利用超声波检测方法对芒果样本进行扫描,将得到的回波信号通过小波变换分析法提取特征向量作为神经网络的输入;构建了一个3层BP神经网络对回波特征向量进行识别和分类。实验显示,该网络用于芒果分级具有较高的准确率。
Feature vector act as neural network inputs,which extracted from the ultrasonic echo signal with wavelet transform analysis.A BP Neural Network with 3 layers is built for the feature vector recognizing and grading.Experiment shows,the trainned Neural Network could be used in mango granding with a good accuracy.
出处
《工业控制计算机》
2016年第10期52-52,55,共2页
Industrial Control Computer
