摘要： 虚拟娱乐照相领域的图像分割,目的在于实时地提取显示出背景中的感兴趣目标,但由于常用的分割方法容易丢失图像的边缘细节信息,而且运算时间长,难以实现质量与速度的平衡,为此文章提出了一种基于遗传量子进化算法的图像分割方法。该方法首先运用遗传量子方法进行阈值搜索,然后进行区域分割,再利用形态学方法对图像边缘细节进行补偿,并去除噪声。仿真结果表明,与传统方法比较,该方法能有效地提取图像中的人像,保留细节边缘。

关键词: 遗传算法, 量子计算, 图像分割, 自适应

Abstract: In virtual entertainment photographing system,goal of image segmentation is to extract and show the interesting part in background.However,using traditional segment methods,some useful edge information may be lost and it needs long time to end,not say to balance the quality and speed.This paper proposes a new method to extract based on genetic quantum algorithm.First genetic quantum method is used to search threshold,then decomposes source image into several areas using quad-tree method,finally,morphological operators are used to compensate edge detail information,and to remove the noise.Simulations show that this method can efficiently extract human part from source image,and reserve detail edges.

中图分类号: 

• TP391.41