神经网络对水稻种子颖壳闭合的识别研究

Research on Neural Network in the Recognition of Rice Seed Hull Closing Level

基于神经网络的基本原理,利用种子的七个特征参数作为输入层,三种类型(正常、微裂、裂颖)种子作为输出层,设计出识别水稻种子颖壳闭合的网络结构。通过大量实验研究表明该神经网络结构能够很好地识别出裂开程度不同类型的种子。

Based on the basic principles of neural network,by using seeds of seven feature parameters as input layer,three types（normal,micro-crack,broken glumes） seeds as the output layer,a network structure was designed to spot the chaff close of rice seeds.A large number of experimental studies had shown that： the neural network structure was able to identify a good degree of splitting different types of seeds.