基于Freeman链码的汉字图像轮廓曲线拐角点检测方法

拐角点检测在模式识别和机器视觉中有重要作用。本文应用基于Freeman链码的一种检测和确定图像轮廓曲线拐角点位置的方法，提取出汉字轮廓图像的拐角点作为特征点用于图像配准中。首先提取出汉字图像的轮廓曲线，然后利用Freeman链码给轮廓曲线编码，根据判断准则，提取出必是拐角点的点与可疑拐角点，最后估算可疑拐角点的曲率，挑出每组临近可能拐角点中曲率最大的点为真实拐角点。实验结果表明本文应用的方法能准确地检测出同被测对象相一致的拐角点，给出了实验数据与过程。

Detection of the Contour Corner of Character Image Based on the Freeman Chain Code

Corner detection plays a very important role in Pattern Recognition and Computer Vision. This paper presents a method for detecting and locating corners of the contour of image based on the Freeman code. The method can extract the corners of the contour of character image as the feature points in Image Registration. The points that are surely the corner points and the doubtful comer points are extracted first. Then the curvature of the points is estimated and the points with maximum curvatures in a set of close points are located as the true corners. Experimental results are provided.