Feature extraction method for round image

The invention relates to a feature extraction method for a round image. The feature extraction method for the round image is characterized in that a plurality of concentric circles are found in a normalized gray level round image, enough points are found at equal intervals on each concentric circle, and the pixels of the points are subjected to binarization processing and are put in a matrix to form a step matrix. Through a linear interpolation algorithm, the row step matrix is subjected to row-to-row stretching transformation. After processing is finished, each line of binarization pixels of the matrix is accumulated to form an accumulated sum curve, a line sequence in the matrix is regulated according to a certain rule, and finally, a binarization feature matrix is formed. 本发明涉及本发明涉及种圆形图像的特征提取方法。本发明所述圆形图像的特征提取方法，在归化后的灰度圆形图像内，找出多个同心圆，在每个同心圆上等间距找到足够多的点，将这些点的像素二值化处理后放到矩阵中，形成行阶梯矩阵。通过线性插值算法，对行阶梯矩阵逐行拉伸变换。处理完成后，对矩阵每列二值化像素累加，形成列累加和曲线，按照定的规则，调整矩阵中列的顺序，最终形成二值化的特征矩阵。