基于偏微分方程图像分割算法研究论文+附录.doc

摘要 图像分割是图像处理中极为关键的一步．图像分割的目的是将图像中我们感兴趣的部分从整幅图像中分离出来，因此分割能力的好坏直接影响了了之后的图像处理效果．自2001年C-V模型被提出以来，基于偏微分方程的图像分割算法开始得以推广．而这也是本文重点研究的方向． 在本论文中，我们首先对传统的图像分割方法进行了简单的阐述并描述了基于偏微分方程的图像分割算法的发展情况．然后重点分析了基于偏微分方程的图像分割算法的基础理论，介绍了在以Kass等人提出的活动轮廓模型的基础上，产生的两个基于区域的几何活动轮廓模型：C-V模型和LBF模型，并解释其工作原理． C-V模型是全局模型，但只能处理二值图像，因此噪音对它的影响极大．LBF模型是C-V模型的改进型，它是局部几何活动轮廓模型，将整块图像先分割成无数小块，通过偏微分方程的计算每个小块得出的值，然后将其相加，它能处理多灰度值的图像包括带噪声的图像． 本文在理论研究的基础上，运用Matlab对涉及到的算法进行仿真实验．通过实验结果来验证本文的各中方法的图像分割效果． 关键词：偏微分方程；图像分割；C-V模型；水平集方法；LBF模型 Abstract Image segmentation is a very important step in image processing. The goal of image segmentation is to segment the target image from the original image, so image segmentation ability directly affects the quality of the image processing results. Since the C-V model was proposed in 2001, the image segmentation algorithm based on partial differential equations has been popularized, which is the key research direction of this thesis. In the paper, we first briefly describe the traditional image segmentation methods and describe the development of image segmentation algorithm based on partial differential equations. And then we analyze the basis theory of image segmentation algorithm based on partial differential equation. Based on the active contour model proposed by Kass et al, the two geometric active contour models based on region are introduced: the C-V model and the LBF model and explain its working principle.. C-V model is a global model, but it can only deal with the two value image. Therefore, the noise has a great influence on it. The LBF model is an improved C-V model, which is a local geometric active contour model. It divide the entire image into numerous small pieces, through the partial differential equations to calculate the value of each small piece, and then add it. It can handle multiple gray values of the image and noising image. Based on the theoretical study, we use Matlab to complete the simulation experiments for the involved algorithms. The image segmentati