摘要： 为了提取彩色图像中线目标，该文提出了一种新的彩色图像分割算法，首先对图像进行分水岭分割得到初始过分割图像，并通过模糊聚类方法得到区域分类概率，然后根据图像的边缘信息和空间特性，得到区域的线方向邻接区域，最后通过迭代的方法，利用线方向邻接区域信息更新区域的分类概率。实验结果显示了很好的提取效果。

关键词: 分水岭分割；LUV 色彩空间；分类概率；线方向

Abstract: In order to extract line objects from color images, this paper presents a new method of color image segmentation. Firstly, original over-segmented image is made by watershed transform and classified probability of each region is obtained using fuzzy clustering method in LUV color space. Then it extracts the neighboring regions in line direction of each region based on edge information and spatial characteristic. Finally the classified probability of each region is updated using the information of its neighboring regions. The experimental results show that the segmentation gives very good extraction result

Key words: Watershed segmentation; LUV color space; Classified probability; Line direction