人工神经网络在猕猴桃种类识别上的应用

Application of Artificial Neural Network for Classification of Gooseberry Species

在猕猴桃介电特性研究的基础上,将人工神经网络技术应用于猕猴桃的种类识别。该种类识别属于模式识别,其关键在于提取样品的特征参数,在获得特征参数的基础上,选取合适的网络通过训练来进行识别。猕猴桃种类识别的研究为自动化识别果品的种类、品种和新鲜等级等提供了一种新方法,为进一步研究果品介电特性与其内在品质的关系提供了一定的理论与实践基础。

Basing on the study of gooseberry＇s dielectric properties, in this paper we make use of the Artificial Neural Network to classify different kinds of gooseberries. This classification belongs to model classification; the crucial point of this classification is getting the characteristic parameters. If we get some kind of characteristic parameters of gooseberries and choose a suitable network, we can classify the gooseberry＇s species by lots of training.