基于自适应量子遗传算法的图像阈值分割

Image threshold segmentation based on AQGA

一维最大类间差法(Otsu)是一种广泛使用的图像阈值分割方法,虽然处理速度快,但是没有考虑到像素的领域空间信息,当图像受到噪声干扰等因素影响时,难以获得满意的分割效果,鉴于此,本文提出了一种基于自适应量子遗传算法和二维Otsu的分割方法。二维Otsu法兼顾了图像的灰度信息以及邻域信息,具有抗干扰以及分割精度高的优点,但存在计算量大、实时性差的缺点;利用自适应量子遗传算法则能迅速找到最佳分割阈值,提高运算速度。仿真实验表明,本文方法减少了阈值寻优的时间,提高了分割精度,同时兼具较强的抗干扰能力。

One-dimensional Otsu method is a widely used image threshold method, although the processing speed is fast, but it do not take spatial information of the pixel into account, while the image is affected by noise and other factors, it is difficult to obtain satisfactory segmentation results, in view of this, a segmentation method based on adaptive quantum genetic algorithm and 2D Otsu is proposed in this paper. The 2D Otsu method considers not only the gray level information but also the neighborhood information, it has the advantages of anti-interference and high segmentation accuracy, but the disadvantages are computationally intensive and poor real-time performance; by using adaptive quantum genetic algorithm, we can quickly find the optimal segmentation threshold, improve processing speed. Simulation results show that this method reduces the time of threshold optimization and improve the accuracy of segmentation, it also has relatively strong anti-interference ability at the same time.