摘要
传统的水平集方法忽略了图像的局部邻域信息,使得水平集曲线易停止于噪声点,导致对含有大量噪声、灰度相近的目标难以分割,且分割结果依赖初始轮廓的选择。因此,本文提出了一种基于马尔科夫随机场(MRF)的自适应水平集图像分割方法。首先利用K-means聚类获得图像的原始先验信息;然后结合MRF获得局部邻域能量信息;最后将MRF能量函数加入水平集中,来约束水平集演化的结果,进而对含有噪声的灰度图像进行自适应分割。通过与一些效果较好的水平集算法进行对比实验,证明了本文方法能够获得更加精确、鲁棒性更好的分割结果。
The traitional level set method ignores the local neighborhood information of the image,so that the horizontal set curve is easy to stop at the noise point,which makes it difficult to segment the target with a large amount of noise and gray scale,and the segmentation result depends on the selection of the initial contour.Therefore,this paper proposes an adaptive level set image segmentation method based on Markov random field(MRF).Firstly,K-means clustering is used to obtain the original prior information of the image;then the local neighborhood energy information is obtained by combining MRF;finally,the MRF energy function is added to the horizontal concentration to constrain the result of the level set evolution,and then the gray image containing noise Perform adaptive segmentation.Through comparison experiments with fuzzy clustering,mean shift and other algorithms,it is proved that the proposed method can obtain more accurate and robust segmentation results.
作者
刘旋
LIU Xuan(Department of Finance,Xinyang Agriculture and Forestry University,Xinyang 464000,China)
出处
《信阳农林学院学报》
2019年第2期99-103,共5页
Journal of Xinyang Agriculture and Forestry University
