基于MATLAB图像分割算法研究与实现.docVIP

基于MATLAB图像分割算法研究与实现 摘 要 图像分割是指把图像分解成各具特性的区域并提取出感兴趣目标的技术和过程，它是计算机视觉领域的一个重要而且基本的问题，分割结果的好坏将直接影响到视觉系统的性能。因此从原理、应用和应用效果的评估上深入研究图像分割技术具有十分重要的意义。本课题主要介绍了图像分割的基本知识，研究了图像分割的两大类算法，即基于边缘检测的方法和基于区域生成的方法。采用MATLAB仿真了所有分割过程，得到了比较理想的分割结果，并分析了各个算法的优点和不足之处,以及适用于何种图像。 基于边缘检测方法种类繁多，主要介绍基于EDGE函数、检测微小结构、四叉树分解和阈值分割的方法实现对图像的边缘检测及提取。而基于区域的图像分割方法主要包括区域生长法和分裂-合并分割方法。通过多次的实验过后，总结出一般的图像分割处理可以用EDGE函数。而特定的图像应用阈值分割、检测微小结构和四叉树分解比较简单。 虽然近年来人们在图像分割方面做了大量的研究工作，但由于尚无通用的分割理论，因此现已提出的分割算法大都是针对具体问题的，并没有一种适合于所有图像的通用的分割算法，有待于进一步解决。 关键字：图像分割；边缘检测；区域生成；阈值分割 Research of image segmentation algorithm Abstract Image Segmentation is the technique and the process to segment an image into different sub-mages with different characters and to extract the interested objects from the image. It is an important and basic procedure in the field of computer vision, the quality of image segmentation directly affects the performance of vision system. Therefore, from the theory, application and evaluation of application effect of depth of image segmentation is of great significance. This issue introduces the basics of image segmentation, image segmentation of the two major algorithms have been done, that is based on edge detection method and the method based on regional produce. Segmentation process is simulated and the results have shown perfect. Advantages and disadvantages of each algorithm are discussed at the end of the paper, and to apply to each image. Edge detection method based on a wide range of EDGE-based functions are introduced, the detection of minimal structure, quadtree decomposition and threshold segmentation method to realize the edge detection and extraction. The region-based image segmentation methods include region growing and division - combined segmentation. Through many experiments later, summed up the general image segmentation can be EDGE function. The specific application of image segmentation, the detection of minimal structure and quadtree decomposition is simple. Although a lot of im