手指静脉图像增强算法研究

The research of finger-vein image enhancement algorithm

根据手指静脉图像的特点,采用了一系列增强图像的算法,分平滑去噪、锐化和二值化三个步骤实现,从而达到分离静脉区域和背景区域的目的。在平滑去噪部分,根据手指静脉图像相邻区域间灰度的关系特点,采用了梯度倒数权重平滑法,对图像进行了去噪处理;在锐化部分,采用将高频强调滤波和直方图均衡化相结合的方法,达到了增强静脉和背景对比度的较理想的效果;在二值化处理部分,根据手指静脉不同区域灰度差别较大的特点,提出了一种分区域处理的方法。试验表明,该算法能有效的分离静脉区域和背景区域。

According to the characteristics of the finger-vein image, we adopted a series o~ methods to enhance the contrast of the image in order to separate the finger-vein areas from the background areas. The method consists of three steps： denoising, contrast enhancement and image binarization. In denoising, considering the relationship between gray levels in the adjacent areas of the finger-vein image, we adopted the gradient inverse weighted smoothing method. In contrast enhancement, we adopted a method which combined the traditional high frequency stress filtering algorithm together with the histogram equalization. With this method, the contrast of the finger-vein area and the background area has been enhanced significantly. During the binarization process, after taking the differences of the gray levels between the different areas of the finger vein image into consideration, we proposed a method which is based on dividing the image into several segments. Our results show that this set of means is quite competent to separate the finger-vein areas from the background areas.