调节形态学运算及其神经网络实现

Regulated morphological operation and its implementation using neural networks

数学形态学是一门建立在集合论基础上的学科,为数字图像处理和分析提供了一种有效的工具。在分析传统的数学形态学基本运算的基础上,引入调节数学形态学运算的概念,然后讨论了调节形态学运算的神经网络实现,并给出了用于图像滤波的计算机仿真结果。该方法较之传统的数学形态学基本运算更为灵活。

Mathematical morphology, which is based on set-theoretic concept, provides an efficient tool to image processing and analysis. Firstly, the basic concepts and properties of regulated mathematical morphology （RMM） are introduced after the traditional mathematical morphological operations are analyzed. Then the regulated morphological operations are implemented by using neural networks. Finally, this method is applied to image filtering. Our experimental results demonstrate that this method is flexible comparing to traditional mathematical morphological operations.