摘要
提出了一种RGB空间聚类和平面区域生长相结合的彩色图像分割方法。聚类时采用重心和空间棋盘距离进行计算,主要为加法、减法运算;区域生长时提出了一种用或连通代替八邻域来判断连通性的新方法,主要为简单的二进制位运算。总体计算复杂度低,从而有效而且快速地实现了彩色图像的目标分割。
This paper advances an algorithm for object segmentation in color image based on RGB space clustering and region growing in plane.Clustering is done mainly with addition and subtraction operations according to grativity center and spatial chessboard distance.While region is growing,a new method mainly with simple binary-bit operation is used to judge connection by OR connection instead of eight neighborhood connection.The algorithm has low computational complexity and makes it possible to segment the object in color image quickly and effectively.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第29期141-143,共3页
Computer Engineering and Applications
关键词
彩色图像
空间聚类
或连通
目标分割
color image
spatial clustering
OR connection
object segmentation
