基于大津法分割和局部最大梯度的自动聚焦算法

Automatic focusing algorithm based on Otsu segmentation and local maximum gradient

针对传统的聚焦评价函数在应用时出现的一些诸如稳定性和精度不足、易受噪声干扰等问题,提出一种基于大津法分割和局部最大梯度的自动聚焦算法。算法对前背景分割后的图像进行局部梯度计算,并统计非零系数和计算局部方差,利用二者变化特性设计一个新的清晰度评价指标以实现数字图像的自动聚焦。实验结果表明,算法具有高灵敏度、较好的稳定性和较强的抗噪能力。

In order to solve the problems of traditional focus evaluation functions such as lack of stability and precision,and vulnerability to noise interference,an automatic focus algorithm based on Otsu method and local maximum gradient is proposed.The algorithm calculates the local gradient of the image after the former background segmentation,and counts the non-zero coefficients and calculates the local variance.Then a new definition value is designed to achieve the automatic focus of the digital image.The experimental results show that the algorithm has high sensitivity,good stability and strong noise resistance.