噪声背景中离散二值图象的形态表示？

本文首先提出了一种新的类型的图象的形态变换，研究了该变换可无失真恢复原始图象的条件及形式，给出了若干重要性质。为了消除变换域中可能存在的噪声，本文提出了基于两种变换的表示方式，并讨论了噪声环境中图象的最优恢复问题，理论分析与计算结果表明该方法比通常的形态骨架变换表示具有更佳的性能。

The Morphological Representation for Discrete and Binary Image Under the Noisy Background

This paper first puts forward a new type of image morphological transformation ,induces the condition under which the original image can be restored with no distortion, and shows several important properties. In order to eliminate the noise possidly existed in the transforming domain,the representation for binary image using two types of transformation is proposed. The optimal restoration in a noisy environment is discussed. Theoretical analysis and computational results have shown this method is better than the usual skeleton transformation.