Abstract: Otsu method is a widely used thresholding segmentation method. To achieve more accurate segmentation of the image, and consider the impact of the edge, a broken-line thresholding segmentation method from the idea that two-dimensional line threshold segmentation instead of the traditional two-dimensional point threshold segmentation is proposed in this paper. The simulation results show that the broken-line thresholding method has a better effect than the original Otsu method, especially for the images with rich edges. The broken-line thresholding segmentation method has better universality than other method in this paper.

Key words: image segmentation, threshold selection, 2-D Otsu method, broken-line thresholding

摘要： Otsu法是一个应用较为广泛的阈值分割方法。为实现图像较为精确的分割，充分考虑边界的影响，从二维线阈值分割替代传统的点阈值分割思想出发，提出了折线阈值型Otsu法。该方法以对边界信息的迭代分割的手段获得实际用于分割的二维折线阈值。仿真结果表明，该方法能够获得优于原始Otsu法的分割效果，特别适用于边缘丰富的图像分割，具有较好的分割普适性。

关键词: 图像分割, 阈值选取, 二维Otsu法, 折线阈值