基于清晰分割的像素覆盖分割方法  

Pixel Coverage Segmentation Method Based on Crisp Segmentation

作　　者：董艳[1] 梁久祯[1] 

机构地区：[1]江南大学物联网工程学院,江苏无锡214122

出　　处：《小型微型计算机系统》2014年第11期2567-2571,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61170121)资助;江苏省研究生培养创新工程(自然科学)项目(CXLX12_0725)资助

摘　　要：提出一种清晰分割方法和像素覆盖分割方法相结合的图像分割方法,估计处于物体边界上的像素的局部覆盖值.由于只有边界上的像素具有部分覆盖,因此提出了一种用清晰分割方法对内部像素进行覆盖值分配,用像素覆盖分割方法对边界像素分配覆盖值的方法,从而减少算法的执行时间.首先使用适当的清晰分割方法对图像进行分割,然后利用模糊形态学边缘检测算法提取不同物体之间的边界,最后采用像素覆盖分割方法对边界像素分配覆盖值.实验结果表明:该方法分割结果优于传统清晰分割方法和FCM聚类算法,可以有效的抑制噪声,减少算法的执行时间.A novel image segmentation algorithm based on crisp segmentation method and pixel coverage segmentation method was proposed. The partial coverage of pixels at object borders was estimated. Because only the boundary pixels were partially covered by different objects, so a crisp segmentation method was used for assigning pixel coverage to pure pixels and a pixel coverage segmenta- tion method was proposed for assigning pixel coverage to mixed pixels. By this strategy, the execution time of the algorithm decreases effectively. First, obtained coverage of pure pixels using a suitable crisp segmentation method. Then, extracted the edges between the different image components by fuzzy morphology for edge detection. In the end,assigned coverage values of edge pixels by the pixel coverage segmentation method. The experimental results show that the algorithm outperforms the result of traditional crisp segmentation method and FCM clustering algorithm, it can effectively eliminate the noise,and the algorithm can effectively reduce execution time.

关 键 词：清晰分割 像素覆盖模型 软分类 模糊分割 模糊形态学边缘检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]