基于顶帽变换的指节纹和掌纹识别改进算法

Improved algorithm of knuckleprint and palmprint recognition based on Top-Hat transformation

针对指节纹、掌纹不易提取,且易受光照和手掌颜色不均影响的问题,提出了基于顶帽变换的指节纹、掌纹识别改进算法。顶帽变换能消除光照不均的影响,很好地提取手掌主纹线。对采集的手掌图像进行灰度化和顶帽变换后,进行二值化处理,分离出手掌区域,找到手掌区域所在位置,用Sobel边缘检测器分离出手掌边界,确定指尖指谷坐标,再对手掌区域进行分割,找到指节纹和掌纹大体区域,用互相关法进行掌纹和指节纹匹配。实验结果表明:提出的方法能够消除光照不均和手掌颜色不均的影响,且速度快,抗噪能力强,适合大规模手掌库实现粗筛选。

Aiming at the problem that knuckleprint and palmprint are not easy to be extracted and the extracting process is easily affected by illumination and fine lines on the palm,an improved algorithm for knuckleprint and palmprint recognition based on the top-hat transformation is proposed.The top-hat transformation can eliminate the influence of uneven illumination and extract palm main line well.Graying and the top hat transformation are carried out on collected palm image,binarization processing is carried out on palm image,isolate the palm area and find its position,and isolate the palm boundary with Sobel edge detector,determine the coordinates of the fingertip and finger valley,segment the palm area,find the knuckleprint and palmprint area in general,match them with the cross-correlation method.The experimental results show that this method can eliminate the influence of uneven illumination and uneven palm color,with fast speed and strong anti-noise capability,suitable for large-scale palm library to achieve coarse screening.