基于形态学连通性理论的图像分割的研究.pdfVIP

第八届中国多智能体系统与控制会议 MASC2012, Oct.20-21,2012, Shanghai，China SatD17.15 基于形态学连通性理论的图像分割研究 1 1 2 1 董天祯 邓廷权 付强 马明华 ( 1、哈尔滨工程大学 模糊信息分析与智能识别研究室 黑龙江,哈尔滨 150001 2、内蒙古通辽市奈曼旗农电局 内蒙古,通辽028300) 摘要：图像分割是图像处理的关键问题，因图像类型和应用领域的不同，尚无通用的、公认的分割方法。 本文运用形态学连通性理论，依据图像的色调、亮度、纹理等分布特性将彩色图像划分成若干个连通区域 (即等价类)。由于连通区域内部像素点具有相同或相近的特征，该方法对平滑区域和纹理部分均有较好的 分割效果。实验证明了本文所提方法的有效性。 关键词：图像分割；形态学连通；图像处理 中图分类号：TP391.41 文献标识码：A Research of Image Segmentation based on Theory of Morphological Connectivity DONG Tian-zhen1 DENG Ting-quan1 FU Qiang2 MA Ming-hua1 (1.Laboratory of Fuzzy Information Analysis and Intelligent Recognition, Harbin Engineering University, Harbin 150001, China 2.Agricultural Electricity Bureau of Naiman District, Tongliao Inner Mongolia, 028300, China ) Abstract: Image segmentatiatin is a key problem in image processing. Owing to various types of images and diversified applications, there is no widely accepted segmentation mehod. In this paper, the morphological connected theory is used to segment a color image into several connected regions(i.e. equivalence class) based on their distributions of hue,intensity and texture, etc. As the pixls in a same connected region posesses the same or similar feature, the proposed method is provided with a sound segmentation result for smooth and texture regions. The validity of proposed method is demonstrated through experiments. Key words: image segmentation; morphological connectivity; image processing; 数学形态学(Mathematical morphology)是一种基 1 引 言 于几何形状和结构思想分析和处理图像的非线性数学