中文摘要：

图像分割在模式识别过程中占有重要地位,如何对图像中感兴趣的目标进行有效的分割是计算机视觉和图像分析的重点和难点.K均值算法在图像分割中具有快速、分割效果好等特点,但其聚类结果易受到初值的影响并且K值的选取也限制了该方法的应用.在基于K均值聚类的基础上,将并行二分的思想引入算法中,并在图像分割中得以应用.通过实验证实,提出的方法在保证图像分割效果的基础上,耗时大大减小.

英文摘要：

Image segmentation occupies an important position in the pattern recognition process,how interested in the image and target for effective segmentation in computer vision and image analysis of the emphasis and difficulties.K clustering algorithm in image segmentation with fast,good segmentation results and other characteristics,but the clustering results was susceptible to the affect of the initial value and K value selection problems limit the application of the method.This paper introduced parallel bisecting concept into the idea of algorithm in image segmentation based on K-means clustering,and used in this field.Confirmed by experiments,the proposed method to ensure the effect of image segmentation,greatly reduced the time greatly.