印鉴识别中图像预处理问题的研究

为便于对印鉴图像进行识别,必须对其进行一系列预处理。针对印鉴图像的特点,阐述了图像预处理的过程,着重研究图像预处理中的杂色去除、灰度化、图像降噪、二值化、空穴填充等。通过在HSV颜色空间中提取红色颜色分量的方法,将印鉴图像提取出来;二值化处理中,采用了结合邻域平均滤波和中值滤波两者优点的新方法。实验表明,经过该方法处理后,可以得到较高质量的图像,处理速度较快,能够满足后期识别的需要。

Image Pre-processing Study for Seal Verification

In order to easily verify seal image, several pre-processing are studied including other clolors removing, noise reduction, binarization, filling in. Useful seal image can be obtained with the method of Withdrawing red component. A new noise reduction algorithm is used which combines the adjacent average algorithm with the middle value filteration algorithm. The method has character of high quality and high speed, establish the good foundation for the next verification.