摘要本文从原理和应用效果上对经典的图像分割方法如边缘检测、阈值分割技术和区域增长等进行了分析。对梯度算法中的Roberts算子、Sobel算子、Prewitt算子、拉普拉斯(Laplacian)算子、LoG(Laplacian-Gauss)算子、坎尼（Canny）算子的分割步骤、分割方式、分割准则相互比较可以看出根据坎尼（Canny）边缘算子的3个准则得出的边缘检测结果最满意。而阈值分割技术的关键在于阈值的确定，只有阈值确定好了才能有效的划分物体与背景，但这种方法只对于那些灰度分布明显，背景与物体差别大的图像的分割效果才明显。区域增长的基本思想是将具有相似性质的像素集合起来构成新区域。与此同时本文还分析了图像分割技术研究的方向。关键词：图像处理图像分割AbstractThisarticleanalysestheapplicationeffecttotheclassicsimagesegmentationmethodliketheedgeexamination,territoryvaluedivisiontechnology,andtheregiongrowthandsoon.ForcomparingtheRobertsoperator,Sobeloperator,Prewittoperator,theoperatorofLaplacianandtheoperatorofLoG(Laplacian-Gauss),Cannyoperatoringradientalgorithm,thestep,thewayandthestandardoftheimagesegmentation,wecanfindoutthethreestandardofCannyedgeoperatortheedgedetectionresultofreachingmostsatisfy.Andthekeypointofthresholdsegmentationlieinfixingthethresholdvalue,itisgoodtohaveonlythresholdvaluetodetermineitthencanbeeffectivetodivideobjectandbackground,butthiskindofmethodisgoodtothosegrayscales,thebigdifferenceimageeffectbetweenthebackgroundandobiect.Thebasicideaofareaistoformthenewregionfromsimilarnature.Andalso,thispaperanalysestheresearchdirectionofimagesegmentationtechnologyatthesametime.Keywords:imageprocessingimagesegmentationoperator1基于MATLAB的图像分割算法研究目录（一般目录要求最多是三级目录，不要出现四级目录）第一章绪论........................................................................................................................................11.1数字图像处理的基本特点........................................................................................................11.1.1数字图像处理的信息大多是二维信息，处理信息量很大（三级标题有问题）.11.1.2数字图像处理占用的频带较宽..........................................................................................21.1.3数字图像中各个像素是不独立的，其相关性大..........................................................21.1.4作合适的假定或附加新的测量..........................................................................................21.1.5数字图像处理后的图像受人的因素影响较大...............................................................21.2数字图像处理的优点.................................................................................................................21.2.1再现性好...................................................................................................................................21.2.2处理精度高..............................................................................................................................31.2.3适用面宽...................................................................................................................................31.2.4灵活性高...................................................................................................................................31.3数字图像处理的应用.................................................................................................................41.3.1航天和航空技术方面的应用..............................................................
