中文摘要：

生菜图像目标分割是基于图像处理的生菜生理信息无损检测的前提。为了解决因生菜富含水分使得图像采集镜头反光而导致生菜叶片图像灰度分布不均的问题,该文采用一种修正的图像灰度均衡算法对生菜图像进行灰度均衡处理,应用混合模糊类间分离聚类算法（MFICSC）进行生菜图像目标分割,使总体类间距离最大化,能够同时生成模糊隶属度和典型值,对处理噪声数据和克服一致性聚类问题均表现良好。分别采用MFICSC算法和Otsu算法进行了生菜图像目标分割对比试验,结果表明MFICSC算法具有较好的聚类准确度,效果优于传统Otsu分割算法。

英文摘要：

Lettuce image target segmentation is the premise of the nondestructive detection of lettuce physiological information based on image processing.Because lettuce contains more water,the camera len is likely to occur reflex,leading to uneven gray distribution of lettuce leaf image.A modified image equalization algorithm is used to equalize the image gray.In this paper,the mixed fuzzy inter-cluster separation clustering（MFICSC） is applied in lettuce image target segmentation,which can make the distance between classes be maximum on the whole and can produce the fuzzy memberships and possibilities simultaneously.MFICSC can overcome the noise sensitivity and the coincident clusters problem.In the test,the MFICSC algorithm and Otsu algorithm were applied to lettuce image target segmentation respectively.The test results show that the MFICSC algorithm has better clustering accuracy,and its segmentation effect is superior to the one of traditional Otsu algorithm