中文摘要：

为实现香梨颜色的自动化分级,建立了机器视觉实验台,将实验台采集到的香梨图像由RGB格式转化为HSI格式进行描述,并采用H分量信息对图像进行背景分割,提取香梨表面像素的颜色信息,建立BP网络对图像颜色进行分类,网络输入层有2个节点,隐含层有7个节点,输出层有1个节点,颜色分类结果的准确率为95%。

英文摘要：

In order to achieve the automation of fragrant pear color grading,a machine vision experiment platform was established.The RGB images of fragrant pears were acquired from the experiment platform,and then transformed the images of fragrant pears from RGB to HIS.The information of H element was used to separate the fragrant pear from background.The color information of the pixel on the surfaces of fragrant pears was counted.A BP artificial neural networks was established to classify the fragrant pears＇ color.There are two nodes of chroma in the input layer,seven nodes in the implicit layer and one node in the output layer.The accuracy rate of classification for color is 95%.