具有鲁棒性的图像骨架提取方法

基于最大正方形的中轴变换,提出了一种二值图像的骨架提取算法。首先通过计算目标像素邻域内最大正方形的所有像素的个数来进行中轴变换,判断骨架点,并提取骨架。如果通过设定一个阈值,再提取骨架,就可以滤除背景中的噪声。实验结果表明:该方法提取的骨架效果好,通过与基于传统中轴变换和形态学提取骨架方法进行比较,该方法抗干扰能力强。

New robust extraction method of image skeleton

Based on the medial axis transform of the upright maximal squares, a new robust binary image skeletonization was proposed in this paper. By computing the numbers of the maximal neighboring pixels in an object, the transform image was obtained, and then skeleton was extracted. If the threshold was initialized, this method could extract the skeleton from the noisy image. Experimental results show that this method can excellently extract the skeletons. Moreover, compared with the traditional medial axis transform (MAT) method and morphology method to skeletonization, this method is robust to background noise.