Abstract: Vein Pattern Recognition is a new contactless biometric technology using IR.Extract vein skeleton with little distortion by thinning the vein pattern separated from the original image is very important to the improvement of successful identification rate.This paper proposes a algorithm to segment，smoothly thin and clip vein pattern based on existing algorithms and analyses of experimentation.Firstly，Gaussian lowpass filter and median filter are used to reduce speck noise and horizontal streak scanning noise in the normalized image respectively，and then the improved NiBlack method and thresholding method based on area are adopted to extract the vein pattern.Subsequently，opening，closing and median filter are applied to smooth the vein edges.Opening and closing are two kinds of morphological operation.Lastly，the improved conditional thinning algorithm and clipping algorithm are employed，and ultimately vein skeletons with little distortion is obtained.The effect of segmentation and thinning is actually improved by the proposed algorithm.

Key words: Vein Pattern Recognition, segmentation, filter, smooth, thining, clip

摘要： 静脉识别技术是一种新兴的非接触式红外生物特征识别技术，从原始图中分割出静脉纹路并细化成变形较小的静脉骨架，对提高识别的成功率非常重要。在现有算法研究和实验分析的基础上，提出了一种分割、平滑细化和毛刺修剪算法，算法对归一化的静脉图像首先采用高斯低通滤波、中值滤波处理，有效地去除了斑点噪声和水平条状扫描噪声，然后用修正的NiBlack局部动态阈值法和面积阈值法分割出静脉，接着对其进行形态学开操作、闭操作及中值滤波以平滑边缘，最后通过条件细化和毛刺修剪得到了失真较小的静脉骨架，获得了良好的分割和细化结果。

关键词: 静脉识别, 分割, 滤波, 平滑, 细化, 修剪