数学形态学运算的再定义及其神经网络实现

Mathematical Morphology Degree-operation and its Implementation Using Neural Networks

该文在分析了传统的数学形态学基本运算的基础上,利用集合势的概念,提出了一种新的形态学算子定义方法,并据此引入了程度化数学形态学运算的概念,将腐蚀和膨胀运算合二为一。最后讨论了程度化形态学运算的神经网络实现及计算机仿真结果,从而,在不同于前人思路的情况下,实现了使用较少的结构元素信息对图象进行噪声滤波的操作。

First this paper addressed an equivalence definition based on the card-inal number of set after analyzing the traditional mathematical morphology operation.Then the erosion and dilation operation were tied into morphological degree-operation.Last in this paper we implemented the above operation using neural networks and simulated the result through computer.So the paper implemented the operation of image filtering by less structure element information.It is an idea different from the previous one.