一种基于遗传算法的灰度图二值化方法

A Binary-conversion Method of Gray Images Based on Genetic Algorithm

遗传算法是当前许多科学实验领域广泛应用的一种非线性并行算法.本文研究了遗传算法在数字图像的灰度图二值化中的应用,提出了一种新的灰度图二值化方法.该方法通过对每个子群体的优化计算和动态改进的适应度函数,确定新的分割阈值.实验验证该方法对于噪声干扰的一般质量图像有着良好的效果.

Genetic algorithm, a kind of nonlinear parallel algorithm, now is widely used in more and more fields, not only computer science but also other science basic. In this paper, a new use of this method will be shown. That is a binary-conversion method of gray images based on genetic algorithm. Based on every process of crossover, replication and mutation, the fitness function modification is-executed depending on the gray center of aim class and center of the background class. When the evolution process is over, a global threshold is defined. Then every image element will be compared with the threshold and made binary according to it. Many examples prove that the method is more effective than the common methods with the gray images disturbed by noise.