基于Gabor滤波和Scharr算法的手掌静脉图像增强方法

Palm vein enhancement based on gabor filter and scharr algorithm

目的为了改善采集到的手掌静脉图像质量,提高静脉识别的准确性,提出一种基于Gabor滤波和Scharr边缘检测算子的手掌静脉图像增强方法。方法首先对原始手掌静脉图像截取感兴趣区域(region of interest,ROI)后进行滤波去噪,其次通过自适应直方图均衡化提高图像对比度,然后采用6个方向的Gabor滤波器提取手掌静脉纹路,并通过Scharr算子实现静脉纹路增强。采用全参考和无参考相结合的图像定量质量评价指标峰值信噪比(peak signal to noise ratio,PSNR)、均方误差(mean squared error,MSE)、灰度标准差、图像清晰度(no reference structural sharpness,NRSS)进行评估。结果本文方法PSNR为29.460,MSN为73.619,图像灰度标准差87.715、NRSS为0.341,结果表明图像失真程度最小,静脉纹路得到有效增强。结论本文方法能够有效增强手掌静脉纹路的对比度,改善图像质量。

Objective In order to improve the quality of the collected palm vein image and the accuracy of feature recognition,we proposed a palm vein image enhancement method based on Gabor filter and Scharr edge detection operator.Methods Firstly,the ROI region of the original palm-vein image was intercepted and de-noised by filtering.Secondly,the image contrast was improved by adaptive histogram equalization.Then,Gabor filter in six directions was used to extract palm-vein lines,and the veins were enhanced by Scharr operator.The quantitative quality evaluation indexes with full reference and no reference,including peak signal to noise ratio(PSNR),mean squared error(MSE),grayscale standard deviation,no reference structural sharpness(NRSS),were used to evaluate the image.Results PSNR was 29.460,MSN was 73.619,grayscale standard deviation was 87.715,NRSS was 0.341.The data showed that the distortion degree of the image was minimal and the veins were effectively enhanced.Conclusions This method can effectively enhance the contrast of palm veins and improve the image quality.