数学形态学的细化算法

Thinning Algorithm Based on Mathematical Morphology

对目标图象进行细化处理，提取其骨架，在目标检测、图象编码以及光学字符识别等计算机视觉、图象处理与模式识别领域都有着广泛的应用．在定义了数学形态学图象轮廓和骨架的基础上，提出了基于两组结构元素模板的数学形态学细化新算法，这两组结构元素模板分别用来去掉目标的西北、东北、东南、西南方向四个角上的点和北、东、南、西四个方向上的点．用该算法对实际的字符进行细化，获得了令人满意的结果．

Object thinning is an important application about mathematical morphology. The methods by which an object is thinned and a skeleton is extracted have been applied to computer vision, image processing and pattern recognition such as object detection, image code, character recognition et al . Based on the definition of contour and skeleton of an object, a new thinning algorithm using two groups of models of structuring element based on mathematical morphology is proposed. By using the two groups of models of structuring element, these pixels are stripped separately in these directions such as northwest,northeast, southeast, southwest, north, east, south and west. The experimental results of the real thinned characters by this thinning algorithm are exciting.