二维熵取小法和模糊熵相结合的图像分割方法

基于二维熵的分割方法是常用的阈值分割技术,其基本假设是对象区域和背景区域占据了二维直方图的绝大部分区域,即假设对象区域和背景区域的概率和近似为1.Brink提出了通过将对象区域的熵和背景区域的熵先取小然后再取大的方法来获得阈值,该方法存在的不足是忽略了边界区域的信息对分割结果的影响,鉴于此,提出了一种结合二维熵和模糊熵的图像分割方法,先采用Brink提出的二维熵法对图像进行初步分割,再采用模糊熵作后续处理以弥补忽略边界信息带来的问题.实验结果表明,对于含噪图像,该方法的分割效果是比较理想的.

Image segmentation method combining two-dimensional entropic method and fuzzy entropy

2-dimensional(2-D) entropic thresholding method is a common image segmentation method.A basic assumption of this method is that regions of object and background cover almost all of the 2-D histogram,that is,the probability sum for object region and background region is close to 1.Brink proposed that the best threshold point is found by maximizing the lesser one between the entropy of background region and that of object region.One drawback is that the method neglects the effect of edges information.In the light of this,an image segmentation method combining 2-D entropic and fuzzy entropy is presented.Firstly,pre-segmentation is made by Brink's 2-D entropic method,secondly,further processing using fuzzy entropy is provided to offset neglecting the information of edges.Experimental results show that the proposed method can obtain better segmentation effect for images with noise.