苹果表面伤痕与果梗和花萼的辨识及检出方法

An Image Recognition and Detection Algorithm for Defect, Stem and Calyx of Fruit Surface

针对在水果表面品质检测的图像处理中,苹果的果梗和花萼很容易被错误辨识成缺陷,提出一种自动识别出苹果缺陷,并计算其大小的图像处理算法,如图像分割、图像去噪、区域标记与提取、感兴趣区域提取等,设计出一种能将采集到的苹果图片进行正确处理后,识别出果梗和花萼,得到真正缺陷大小的图像处理算法。运用C++语言进行编程,采用MFC基于对话框的非模态形式设计了图像处理软件,处理采集到的苹果样品图片,结果表明,该算法能正确计算缺陷大小,并且能得到果梗和花萼的辨识依据,可用于水果表面品质实时检测中。

In computer image processing of apples, the stem and calyx are usually mistaken as defect, so a useful algorithm that can automatically identify the defect and calculate its size is proposed. Methods such as image segmentation, image denoising, region labeling, and ROI extraction are applied in theoretical analysis and practical programming. Ultimately, an image processing algorithm applicable to the subject is designed that can correctly identify stem and calyx of apple samples and get the size of real defect. The image processing software is devised by MFC based on the non-model state form of dialog. Moreover, the system utilized C＋＋ language to programming. The results show that the size of defects can be calculated effectively and get the identification basis of item and calyx. And it can be used for real-time detection of fruit surface quality.