二维直方图区域斜分阈值分割及快速递推算法

Image thresholding based on two-dimensional histogram oblique segmentation and its fast recurring algorithm

指出了二维直方图区域直分法中存在明显的错分,提出了一种新的二维直方图区域斜分方法,即通过4条平行斜线将直方图分成内点区、边界点区和噪声点区,并采用与主对角线垂直的斜线进行阈值分割;然后导出了基于二维直方图区域斜分阈值选取的公式及其快速递推算法;最后给出了分割结果和运行时间,并与Otsu快速算法进行比较。结果表明二维直方图区域斜分方法可以运用于几乎所有的基于二维直方图的阈值分割,使分割后的图像内部均匀,边界准确,抗噪更稳健,同时其运行时间大幅减少。

The obvious wrong segmentation in the existing two-dimensional histogram vertical segmentation method was pointed out. A new two-dimensional histogram oblique segmentation method was proposed, in which the histogram was divided into inner, edge and noise parts by four paralleled oblique lines. The image was thresholded by the oblique line which was perpendicular to the main diagonal. Then the formulae and its fast recurring algorithm of the two-dimensional histogram oblique segmentation method were deduced. Finally, the thresholded images and processing time were given in the experimental result and analysis, which were compared with those of Otsu fast algorithms. The results show that the two-dimensional histogram oblique segmentation could be used in nearly all the two-dimensional histogram thresholding. It makes the inner part uniform, the edge accurate in the thresholded image and has better tolerance capability to noise. At the same time the running time is significantly reduced.