中文摘要：

以新疆库尔勒香梨为研究对象,提出使用核磁共振成像技术检测香梨内部褐变的方法.采用核磁共振成像设备扫描得到香梨中部冠状面图像,使用MATLAB软件完成图像分割和中值滤波,采用自动阈值分割目标区域,利用图像面积区域属性函数分离果核与褐变部分,以提取香梨褐变特征,并对误判褐变切片图像作了进一步分析与处理.实验结果表明：核磁共振成像技术对不同贮藏周期的正常与褐变香梨切片图像的识别准确率分别为84%和100%,该方法对贮藏后期（6个月）香梨的褐变识别更有效.

英文摘要：

By adopting Korla fragrant pears （Pyrus bretehezderi Rehd. ） as the experimental object, the nuclear magnetic resonance imaging （MRI） technology was proposed to detect the internal browning on fragrant pears. Coronal magnetic resonance images were acquired by horizontal scanning mode in the middle of pear. Image segmentation and median filter were finished by MATLAB software. Otsu~s threshold was used to segment object image. The feature of browning was extracted through the area property of image region which can distinguish browning with the kernel of pear. The misclassified browning slice images were further analyzed and reprocessing. The experimental results showed that detecting browning on pears with MRI was realized and the accuracy was 84% and 100% for sound and internal browning pears at different storage time respectively. The method for fragrant pear browning recognition with longer storage time（6 months） was more effective.