Abstract: Focusing on the application on the offline embedded system, a fast algorithm for the large area rolling fingerprint’s capture and mosaicking is proposed. After the first matching using the block direction characters got from the enhanced fingerprint image, the second matching has been executed based on texture information to get better matching position. The curve stitching has been used to realize fingerprint mosaicking. Experimental results show that, the algorithm has good precision, robustness, adaptability, portability and fast speed, low resource using, so it can be well used in real-time system.

Key words: fingerprint mosaicking, rolling fingerprint, block direction, texture information, twice matching, curve stitching

摘要： 针对大面积滚动指纹采集和拼接在离线嵌入式系统中的应用，提出一种快速算法。对采集到的指纹图像进行增强处理，提取指纹块方向特征进行初次匹配，在此结果上利用纹理信息进行二次求精匹配，采用曲线缝合技术进行缝合。实验结果表明，该算法具有速度快、精度高、资源耗费小、鲁棒性好、适应性强、移植方便的特点，可满足实时嵌入式应用。

关键词: 指纹拼接, 滚动指纹, 块方向, 纹理信息, 二次匹配, 曲线缝合