基于分块的图像分割方法研究

图像分割是图像分析的关键步骤,基于阈值的图像分割方法在各领域得到了广泛的应用。在分析图像分割相关背景理论的基础上,针对固定全局阈值在图像分割中的不足,提出了一个基于分块的图像分割方法。首先将待处理图像分为若干大小相同的块,减少被处理对象所含的异常;然后采用基于矩的阈值的选取方法,分别对每一块图像数据进行二值分割处理;最后将各块数据汇总,重新恢复成一幅图像。实验证明了该方法在人物背景分离、基本轮廓提取等方面的有效性。

Research on Block- Division Based Image Segmentation Method

Image segmentation is one of the key steps in image analysis, where the threshold based methods are widely used. After a brief introduction of image segmentation, a block-division based segmentation method is raised in this paper, which overcomes the shortcomings of the whole and single threshold method. The image is firstly divided into several blocks with the same size, reducing the abnormitis in one block; then the quadrature based method is adopted for threshold selection; each block is processed with two-value segmentation method; and finally, all blocks are combined for image recovery. Experimental evidences show the validity of proposed method in background separating and contour distilling.