基于二维OTSU选取种子点的区域生长图像分割  

Image Segmentation Using Region Growing Based on 2D OTSU to Select Seed Points

作　　者：吴海滨[1] 周雨润[1] 周英蔚[1] 陈新兵[1] 项龙飞[1] 李梓霖 

机构地区：[1]安徽大学物理与材料科学学院,安徽合肥230601

出　　处：《大气与环境光学学报》2013年第6期448-453,共6页Journal of Atmospheric and Environmental Optics

基　　金：科技部创新基金项目(07C26213400516)资助

摘　　要：在对连通性要求比较高的图像进行分割时,区域生长方法有独特的优势。在图像复杂的情况下,希望计算机可以快速并且完整地自动选取种子点。利用二维OTSU对图像进行预分割并以此作为选取种子点的依据,可以有效解决使用区域生长分割图像时种子点选取的难题。结合对石化炼油炉炉管图像的分割,讨论使用此种方法的可行性和有效性。For image segmentation in which connectivity requirements are relatively high, region growing method has unique advantages. In the complex situtaion, that the computer can fast, completely and au- tomatically select seed points is needed. Using two-dimensional OTSU pre-segmentation of the image can effectively bined the method is solve the problem of seed points selection by using region growing image segmentation. Com- image segmentation with petrochemical refining furnace tube, the feasibility and effectiveness of this discussed.

关 键 词：区域生长 种子点 二维OTSU 预分割 结合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]