摘要图像处理在它漫长的研究历程中已经慢慢地形成了一套自己的科学体系，所以它的新处理方法也随之而层出不穷。而在这个图像处理发展的过程中，图像分割又一直都是最基础，也是最重要的领域之一。所以不仅得到国内人士的关注和研究也受到了海外研究人士的关注，也在很多方面得到应用并取得一定的成果，但是也一直都有一个缺憾就是到目前为止还没有人能研究出一种唯一的标准分割法，所以在未来的发展历程中还需要我们多加探索，挖掘图像分割更多的方法【1】。80801

    图像分析的那些经典的研究课题就包括了灰度阈值分割、边缘检测和区域法。就现在已经研究出的理论和方法都有很多不到位的地方，所以我们仍然需要不断的对这些算法进行改进和进一步的发展。我们现在所常用到的一些经典分割运算都能对任何图像进行一个直接的操作，但是仍然有些算法应用范围比较狭窄，只能针对有一些比较特殊的图像。为了更精确的从图像中获取更多我们想要的信息，我们不得不在面对某些图像处理前先进行一个粗分割的操作流程【2】。另外，我还要强调说明的是边缘跟边界并不是同一个意思，它们所代表的意义也不一样。如果一个图像中的某个地方的像素值存在着突变，那么这个地方就是边缘；而边界则位于我们已知存在的物体跟生活的真实场景之间【3】。除此以外，在成像的过程中需要我们特别考虑到的两大因素就是光照和噪声。由于现实中的图像都存在特殊性跟不规则性，所以有很多分的图像分割都没有办法完全依靠计算机来自动完成，但是如果人为进行操作分割的话，又存在着很多的定位不准确跟操作量太大的问题。因此，就有人提出了采用人跟计算机相互配合进行操作的想法，利用计算机的定位精准跟人细节上的把握从而导致对目标轮廓的一个更快更精准的定位，既省时又省力【4】。若上述的人工交互式能得到实现，那么这个具有重要学术价值的课题——图像分割，一定具有广阔的应用前景，且被越来越多的人加以研究和创新。

毕业论文关键词：阈值分割；边缘检测；区域法

Abstract Image segmentation is one of the most basic and important fields in image processing。 And the image processing and analysis itself with the development of science and technology has gradually formed its own scientific system, the new processing methods are emerging。 Image segmentation and edge detection in the past twenty years get the wide attention and rapid development, domestic and foreign many researchers proposed many methods, has made some achievements in different fields, but so far has not only the Standard segmentation methods,so there are a lot of exploring space。

    Image analysis of the classic research topics include the gray threshold segmentation, edge detection and region method。 But there are many deficiencies in the theories and methods of the present research, which still need our continuous improvement and development。 Most of the segmentation operations can be directly applied to any image, only a small portion of the image can only be applied to some special。 Some images in the image segmentation before we need to make a rough segmentation, because the need to extract information from the image。 In addition, I also need to explain the point is, edges and boundary of an object is not the same, the edge of the meaning is image pixel value where mutations, and refers to the object boundary is the boundary between the existence of the object and the reality of the scene。 In addition, the two factors that need to be considered in the process of imaging are light and noise。 Due to the image of the significance and complexity, many segmentation can not rely on computer automatically, and manual and workload, positioning inaccurate problem。 So, it was proposed that the combination of some human interaction and computer automatic positioning method, take advantage of to achieve fast positioning of the target contour。I believe these interactive method if it can be fully applied,this has important academic value of the subject,image segmentation,it must have broad prospects for application and further study,also will become a more active field of research。  MATLAB的图像分割算法分析与仿真实现:http://www.youerw.com/zidonghua/lunwen_94040.html