中文摘要：

果梗是香梨品质分级的重要指标之一。本文在香梨图像经过背景分割和边缘检测获取边缘图像后,提出一种利用数学形态学方法对香梨果梗进行自动提取的新方法。利用MATLAB 7.1软件平台开发了香梨果梗自动提取的计算机视觉检测软件。研究结果表明,该方法对香梨果梗的提取的正确率为90.7%。

英文摘要：

Stem is one of the important index to classify fragrant pears according to their quality.After the optimal edges were acquired by filtering and edge detection,a new detection method that automatically extract the stem of fragrant pear was proposed based on the Mathematical Morphology.A Kurle fragrant pear stem detection extraction software was developed based on the platform of MATLAB 7.1.Experimental analysis results show that the accurate rate of extracted stem reached 90.7% by the method.