Abstract: Image segmentation is an important technology in the image processing，pattern recognition and computer vision area.In order to obtain high-quality image segmentation，it proposes a novel method for digital image segmentation based on gray equalization and improved genetic algorithm.On the one hand，this method combines the genetic algorithm and Otsu，and improves the performance of genetic algorithm;on the other hand，it uses the gray equalization to enhance the image before the solving process of genetic algorithm.From the experiments，the method can overcome the problem which is the common image threshold segmentation methods in preserving the detail of gray image.It can obtain the best threshold value stably and achieve the high quality image segmentation effects.

Key words: image segmentation, genetic algorithm, Otsu, gray equalization

摘要： 图像分割是图像处理、模式识别、计算机视觉等领域的重要技术。为实现高质量的数字图像分割，提出了一种结合图像灰度均衡和改进遗传算法的数字图像阈值分割方法。创新点在于一方面采用结合了最大类间方差法的改进遗传算法，并对遗传算法性能加以改进;另一方面，对图像进行了灰度均衡的图像前处理，使得算法具有更广泛的适应性。实验显示，方法克服了常见的图像阈值分割方法在处理灰度图像时出现的图像细节难以保留的问题，能够稳定地获得图像的最优阈值，实现保留图像细节的分割效果。

关键词: 图像分割, 遗传算法, 最大类间方差, 灰度均衡