手写数字字符的边界特征提取

Boundary feature extraction of handwritten numeric characters

为提高图像质量,研究手写数字字符的边界特征提取技术.设计了一种图像预处理方案,给出了灰度化和二值化的具体算法公式,主要运用Roberts算子、维纳滤波和形态学处理三种算法成功提取出了手写数字字符的边界特征,字符边界清晰,内部饱满无空缺.该图像预处理方案为正确分割字符打下了良好的基础.

In order to improve the image quality,we studied the boundary feature extraction technology of handwritten numeric characters.We designed an image preprocessing scheme,gave the specific algorithm formulas of grayscale and binarization,and successfully extracted the boundary features of handwritten numeric characters by using the three algorithms:Roberts operator,Wiener filtering and morphology processing.It turns out that the character boundary is clear,and the interior is full without vacancy.The image preprocessing scheme lays a good foundation for the correct segmentation of characters.