基于样条插值的图像分割算法研究

Image Segmentation Algorithm Study Based on Spline Interpolation

针对传统分割算法中的阈值选择问题,提出了一种改进的图像阈值分割算法。采用样条插值方法对图像的灰度直方图进行曲线拟合,获得直方图的曲线表达。利用函数极值的特点寻找直方图曲线的极值点,获取阈值,对图像进行阈值分割。实验结果表明,改进算法获得的阈值较其他算法有更好的分割效果。

In order to solve the threshold selection problem of the traditional segmentation algorithms, an improved spline interpolation algorithm for image segmentation is presented. Firstly, the spline interpolation is used to obtain the image histogram fitting curve. Secondly, the extreme points of the histogram curve is found based on the characteristic of the function extreme and the threshold is obtained. Then, the image is segmented with the threshold. Numerical experiments show that the threshold segmentation obtained by the improved al-gorithm is better than other algorithms.