摘要
如何去除毛刺是图像处理中一个经典的问题和研究热点,目前有些方法只能去除较短的毛刺,不能适合特定的需要,有些方法存在部分骨架主体被去除的缺点。对基于模板去除毛刺的方法进行了改进,该方法首先在8邻域模板范围内进行统计,确定相应的像素类型,然后基于人工定义毛刺的长度进行毛刺去除。实验表明,该方法能够适应不同长度的毛刺,并且可以尽可能地保持骨架主体信息,对单像素宽度的骨架图像去除毛刺具有一定的可行性和优越性。
Removing burrs is a classical problem and research pot.At present,some methods can only remove short burrs,and some other methods have the defect of removing the skeleton information.So,an improved template-based method of burr removal was proposed in this paper.Firstly,the method recorded pixels number in the 8 adjacent domain determined and defined pixel type.Then burrs were removed based on the length threshold.The experimental results prove that the method can fit for burrs with different length,and can maintain main body information of skeleton.So,the method for removing skeleton burrs has some feasibility and superiority
出处
《计算机应用》
CSCD
北大核心
2011年第A01期58-59,63,共3页
journal of Computer Applications
关键词
去除毛刺
骨架图像
模板
连通区域
burr removal
skeleton image
template
connectivity zone
