摘要
针对现有并行细化算法中存在的细化不完全、二像素宽斜线细化畸变等问题,提出一种改进的细化算法.首先对二值图像构造了一组保留模板和消除模板,采用算术逻辑运算对图像进行初步细化,并结合保留模板匹配将二像素宽斜线予以保留;再利用消除模板匹配进一步删除斜线上的冗余像素,得到8连接的单像素图像.针对字符图像的实验结果表明,文中算法能够有效地避免二像素宽斜线的细化畸变,保持了原图像的拓扑结构,实现图像的完全细化.
In order to overcome the shortcomings of existing parallel thinning algorithms, such as thinning incompletely and 2-pixels slash thinning distortion, an improved thinning algorithm is proposed. A set of reservation templates and elimination templates for binary image were constructed firstly. The elementary thinning process was performed by arithmetic and logic operation, while 2-pixels slash was reserved by reservation templates matching. Then the redundancy pixels on slash were deleted by elimination templates matching, and the 8-connected single-pixel image was acquired. Experimental results on character images show that, the proposed algorithm could effectively avoid the 2-pixels slash thinning distortion, maintain the topological structure of original image, and achieve complete thinning result.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2013年第1期62-66,共5页
Journal of Computer-Aided Design & Computer Graphics
基金
内蒙古工业大学科研基金(200602)
关键词
细化
并行算法
模板匹配
二像素宽斜线
thinning
parallel algorithm
template matching
2-pixels slash
