【摘要】 文章提出了一种有效的基于颜色和纹理综合特征的图像分割方法。将图像以块为单位进行划分,在YUV空间,提取块的颜色特征和纹理特征,在这种综合特征基础上,采用改进的K均值聚类法进行图像分割。该方法能自适应确定聚类中的参数,且兼顾点的位置连通关系,从而达到了较好的分割效果。更多还原

【Abstract】 An effective method for image segmentation based on color and texture features is proposed.An image is par-titioned into blocks with pixels,and a feature vector which in cludes color and texture features is extracted for each block.In YUV color space,an improved k-means algorithm which can automatically determine parameters and take into account both feature similarity and location neighboring relation is used to cluster the feature vectors into several class-es with every class corresponding to one region in segmented image.Experiments show that image segmentation method in this paper is effective.更多还原