基于特征描述子的匹配方法

摘要

利用一种效率和匹配率均较高且存储空间小的基于特征描述子的指纹识别算法，实现了指纹匹配且可应用到嵌入式系统。由于自动指纹识别系统一般在嵌入式系统上实现，对算法的复杂度、运算速度、存储空间需要等均有较高要求，因此选用普遍使用的基于细节特征的匹配算法。该算法运算速度快，一幅300×300像素的指纹图像，仅需几十个特征向量。实验证明，算法克服了点匹配计算量大、速度慢和参考点确定难等缺点，且对图像的旋转和平移具有较好的鲁棒性。可以使基于特征描述子的随机局部区域描述子特征的匹配性能进一步提高。%A fingerprint identification method based on the has high efficiency and matching and small storage space. By this characteristic descriptor is proposed. This method method fingerprint matching is realized and applied to the embedded system. Because automatic fingerprint identification is usually realized in an embedded system, it has a high requirement on the complexity, calculation speed and storage space of an algorithm. Therefore, a popular characteristic based matching method is used. This algorithm is fast in calculation. Experiments show that this algo- rithm has the advantages of simple calculation, high speed easy determination of the reference point and good robust- ness on the rotation and translation of the image. By this algorithm, the matching of the characteristics of the random local descriptor based on the characteristic descriptor can be improved.