【作者】 王浩；

【导师】 彭宏；

【作者基本信息】 西华大学 ， 计算机软件与理论， 2012， 硕士

【摘要】 膜计算（P系统）是从生物细胞以及由细胞组成的组织和器官的功能和结构中抽象出来的计算模型。P系统的分布式、极大并行性、非确定性以及膜优化算法的较好的适应性、较强的寻优能力对于求解复杂优化问题具有一定的优越性。如何将P系统中这些特征和优势应用于图像处理，对于拓展P系统的应用范围和研究新的图像处理算法都具有重要的意义。本文重点关注于P系统在图像分割上的应用。本文基于组织型P系统的机制和并行计算优势，并结合经典的最大类间方差法（Otsu）和最佳熵法（KSW）原理探讨了基于P系统的单阈值图像分割和多阈值图像分割。所取得的创新成果如下：(1)提出了一种基于组织型P系统的单阈值分割方法。设计了一个两层膜的组织型P系统。结合Otsu准则和KSW准则，通过该P系统的进化规则和转运规则搜索最佳分割阈值。通过与传统阈值方法和基于遗传算法的阈值方法的比较验证了所提出的阈值方法的有效性和可行性。(2)提出了一种基于组织型P系统的多阈值分割方法。依据所开发的组织P系统的机制，并结合多阈值分割原理为图像多阈值分割搜索最佳的多阈值。通过与传统多阈值方法和基于遗传算法的多阈值方法的比较验证了所提出的多阈值方法的有效性和可行性。更多还原

【Abstract】 Membrane computing (P system) abstracts computing models from the architecture and the functioning of living cells as well as from the organization of cells in tissues and organs. P system that possesses distributed, maximum parallel and non-deterministic computing features, and membrane algorithm that has a better adaptability and strong optimization ability, have advantages for solving complex optimization problems. How to apply these features and advantages to image processing is of great significance for both extending the application range of P system and developing new image processing algorithms. This paper focuses on the application of P system to image segmentation.Based on the mechanism of tissue-like P system and its parallel computing advantages, this paper discusses the single-threshold image segmentation and multi-threshold image segmentation by combining classic maximum between-class variance (Otsu), entropy method (KSW) principle and P system. The main innovations are shown as follows:(1) Based on tissue-like P system, a single-threshold based image segmentation method is proposed. We design a tissue-like P system with two layer membranes. Combining Otsu’s criterion and Ksw’s criterion, the P system searches the best segmentation threshold by the evolution rules and communication rules. We compare with traditional thresholding method and thresholding method based on genetic algorithms. The comparison results demonstrate the effectiveness and feasibility of the proposed thresholding method.(2)A multi-threshold image segmentation method based on tissue-like P system is proposed. According to the mechanism of the tissue-like P system and the principle of multi-threshold image segmentation, P system searches the best multi-threshold for multi-threshold image segmentation. By comparing with traditional thresholding method and thresholding method based on genetic algorithms, the effectiveness and feasibility of the proposed thresholding method are verified.更多还原