关于掌纹图像基于形态学的角点检测的定位分割方法

Location Method of Palmprint Image Based on Morphology and Corner Detection

为了更好地进行掌纹识别,掌纹图像的预处理是必不可少的。提取掌纹图像的感兴趣区域(ROI)是掌纹图像预处理的一项重要内容。采用基于形态学的方法,将进行膨胀和腐蚀后的图像相减得到的图像作为掌纹图像的轮廓。根据掌纹图像的轮廓利用Harris算法进行角点检测,在检测出的角点中,求出角点的曲率,将曲率最大的两个角点作为定位的基准点来建立坐标系提取ROI。根据实验结果,该算法取得比较理想的效果。

For the better palm-print recognition, palm-print image pre-processing is essential. It is the important content of taking ROI in palmprint image pre-processing. The method is based on morphology, uses the dilated images to subtract the eroded images as the outline palm print images. According to the outline of palm-print images, uses Harris algorithm for comer detection. In the detected comer, the curvature of the two largest comers as a positioning reference point to establish the coordinate system to extract ROI. According to the resuits, the algorithm achieves a more satisfactory results.