中文摘要：

该文提出了一种基于图像处理的方法检测猕猴桃的损伤和疤痕缺陷检测方法。首先，通过阈值分割，将猕猴桃图像与背景分割。其次，通过分析损伤和疤痕猕猴桃的颜色特征，选用YcbCr颜色模型的Cr（红色）成分对猕猴桃损伤果检测效果良好，选用HSV颜色模型中的V（亮度）成分用于对猕猴桃疤痕果检测效果良好。对50个损伤和疤痕样本的猕猴桃检测成功率达88％，为猕猴桃缺陷无损检测提供了一种新的方法。

英文摘要：

This paper presents a defect detection algorithm for kiwifruit with spot and injure by image processing. First, by using threshold method, kiwifruit image is segmented from the background. Second, after analyzing the color feature of kiwifi＇uit with spot and injure, we select Cr channel of YCbCr color model and V channel of HSV color model to detect defect feature. For 50 samples with 25 spot and 25 injure separately, the proposed method get 88% overall accuracy. It provides a new algorithm for detecting kiwifruit defect by using image processing.