文章以胡柚为研究对象， 针对基于机器视觉的胡柚品质分拣生产线所涉及的关键技术进行理论和试验研

究，先对在传输带上运动的胡柚进行图像采集、图像分割、图像平滑、灰度化和锐化等一系列 的图像处理， 然后对处理后

的胡柚图像提取大小、颜色和缺陷的特征值，最后依据提取的特征参数进行大小、颜色和缺陷分级。 并以此为基础，研究

适合实时条件下的胡柚大小、形状、颜色及果品缺陷等品质指标的检测方法和分拣执行机构。- This paper takes grapefruit as the research object and conducts study on key technologies for quality sort-

ing grapefruit production line based on machine vision. The first step is to conduct the image processing for grapefruit

moving on the transmission belt such as image acquisition, image segmentation, image smoothing, gray and sharpening se-

ries of image processing, and then extracts characteristic values of the size, color and defects　the processed images.

on this basis, it develops defect detection method and sorting actuators under the conditions that suit the real grapefruit

size, shape, color and fruit quality characteristics.