基于自适应Harris角点检测的掌纹定位方法

Fast location algorithm based on adaptive Harris corner detector

分析Harris角点检测算子的实现原理及其应用于掌纹定位中时的不足,提出了一种基于自适应Harris角点检测的快速掌纹定位算法,该算法可以人为选择感兴趣区域进行检测,然后利用角点进行定位。有效地避免了掌纹定位过程中二值化,边缘检测等一系列复杂的过程,并且克服了检测过程中的角点簇拥,计算量大的缺点。实验结果表明,该方法进行的掌纹定位非但计算量小,处理速度快,而且检测出的角点分布均匀合理,能更好的适应图像定位等一系列的应用。

The realization theory of Harris corner detector and its disadvantage in application of palmprint location are analyzed. A fast location algorithm based on adaptive Harris corner detector is proposed. The region interested can be detected artificially using this algorithm and then corner detector is used for location. This method effectively avoids binaryzation and edge detection in palmprint process. Moreover, it has overcome the clustering of corner and the disadvantage of large amount of calculation in the process of detection. The experimental result shows that this method not only has small amount of calculation and fast processing speed, but also the distribution of detected corner is reasonable, so that it can be used in image location.