摘要
针对经典Otsu算法计算量大、实时性与抗噪性差的问题,提出了一种基于最小二乘法拟合的Otsu快速图像分割方法。首先,算法在实验图像有效灰度区间上选取9个均匀分布的灰度点,同时计算对应的类间方差数值;其次,利用最小二乘法对这9个点类间方差数值进行二次曲线拟合;最后对二次曲线二次求导,求取拟合曲线最大值时对应的阈值。实验结果表明,算法显著提高了计算速度与搜素效率,减少计算方差次数。
In view of the problems of large computation amount,poor real-time and antinoise capability of the classical Otsu algorithm,a fast image segmentation algorithm based on least square method is proposed for the Otsu algorithm.Firstly,nine evenly distributed gray points are selected in the effective gray range of the experimental image,and the corresponding between-cluster variance values are calculated.Secondly,the least square method is used to fit the variance between the nine points;finally,the quadratic derivative is used to obtain the threshold value of the maximum value of the fitting curve.The experimental results show that the algorithm can significantly improve the computational speed and search efficiency,and reduce the number of calculation variance.
作者
徐建东
XU Jiandong(Jiangsu Guoguang Information Industry Co.,Ltd.,Changzhou 213001,China)
出处
《常州大学学报（自然科学版）》
CAS
2021年第1期70-76,共7页
Journal of Changzhou University：Natural Science Edition
