基于HSI空间的生长状态下茄子图像的分割方法  

Segmentation Methods of Eggplant Images in the Growth State Based on HSI Space

作　　者：刘健[1] 张云伟[1] 

机构地区：[1]昆明理工大学现代农业工程学院,云南昆明650224

出　　处：《安徽农业科学》2010年第14期7554-7556,共3页Journal of Anhui Agricultural Sciences

摘　　要：茄子识别是茄子采摘机器人视觉系统的关键技术,利用Matlab软件中的图像处理函数,将在自然光照条件下拍摄的大棚中成熟茄子图像进行分割。对H、I通道分别进行自动阈值法和固定阈值法分割,将分割后二值化图像进行逻辑与运算,获得最终分割结果。去除噪音部分,使用形态滤波的开运算消除图像中其他的噪音部分。结果表明,该分割方法正确率大于91%,是一种较实用的茄子分割算法。Eggplant identification is one of the key technologies of vision system of the eggplant picking robot. Using the image processing func-tions in Matlab software,the mature eggplant images in shed under natural light conditions were segmented. H and I components were conducted by auto-threshold value method and a fixed threshold value segmentation. After the segment,the binary images after segmentation were added to obtain the final segmentation results. Upon the elimination of the residue noise,morphological filtering was used to eliminate other noise in the image. The results showed that the accuracy rate of the segmentation method was more than 91% and it was a more practical segmentation algorithm of eggplant.

关 键 词：茄子 采摘机器人 视觉系统 图像分割 

分 类 号：TN911.73[电子电信—通信与信息系统]