计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (16): 160-162.

• 图形、图像、模式识别 • 上一篇    下一篇

基于水平集的并行融合图像分割

刘敬伟,李 岑,侯 帅   

  1. 北京航空航天大学 数学系 数学、信息与行为教育部重点实验室,北京 100083
  • 收稿日期:2007-11-26 修回日期:2008-01-02 出版日期:2008-06-01 发布日期:2008-06-01
  • 通讯作者: 刘敬伟

Image segmentation using parallel fusion method based on level set

LIU Jing-wei,LI Cen,HOU Shuai   

  1. LMIB & Department of Mathematics,Beijing University of Aeronautics and Astronautics,Beijing 100083,China
  • Received:2007-11-26 Revised:2008-01-02 Online:2008-06-01 Published:2008-06-01
  • Contact: LIU Jing-wei

摘要: 针对复杂背景和多目标空中运动物体的定位和跟踪问题,研究了基于小波变换阈值去噪方法、数学形态学去噪方法与水平集方法结合的图像分割方法,提出了基于小波变换阈值去噪与水平集结合方法以及基于小波变换阈值去噪、数学形态学去噪与水平集结合方法的并行融合图像分割方法,对运动目标进行边缘检测。实验结果说明,基于小波变换、数学形态学与水平集方法结合的并行融合图像分割方法能够有效地提取目标物体的轮廓,抑制背景噪声。

关键词: 图像分割, 水平集, 小波变换, 数学形态学, 边缘检测

Abstract: To survey the localizing and tracking of aerial moving mulit-objects in complex background,the image segmentation methods of wavelet threshold de-noising method,mathematic morphology de-noising method and level set method are investigated,and a parallel fusion method of combination of two edge detection methods,the one is combination of wavelet with level set,the other is combing wavelet and mathematic morphology with level set,is proposed to detect the contour of moving objects.The experimental results show that parallel fusion combination method of wavelet,mathematic morphology and level set effectively extracts contour edging of objective and denoises background noise.

Key words: image segmentation, level set, wavelet transform, mathematic morphology, edge detection