基于分形和分水岭的图像分割方法  

IMAGE SEGMENTATION BASED ON FRACTAL THEORY AND WATERSHED SEGMENTATION

作　　者：陈琳[1] 刘文奇[1] 蔡光程[1] 

机构地区：[1]昆明理工大学数学系,云南昆明650093

出　　处：《计算机应用与软件》2008年第8期227-228,共2页Computer Applications and Software

摘　　要：图像分割是一种重要的图像处理技术,也是计算机视觉领域低层次视觉中的主要问题,同时它又是一个经典难题。提出了结合分水岭分割与图像分形维数的一种新方法用于对自然背景下人造目标的提取。实验结果证明,该方法能有效抑制自然背景,并提取出人造目标的轮廓。Image segmentation is an important image processing technique and the main problem in low level vision of computer vision field; it is also a classical poser. A new algorithm was proposed for the extraction of artificial target segmentation with natural background based on image fractal dimension and watershed segmentation. The experiment results show that this algorithm can restrain natural background effectively which lead to an extraction of complete contour of the artificial target.

关 键 词：图像分割 分水岭分割 分形维数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP18[自动化与计算机技术—计算机科学与技术]