一种改进的Sobel梯度函数自动对焦评价算法  

An improved auto-focus evaluating algorithm based on Sobel gradient function

作　　者：张丰收[1,2] 李斯文[1] 胡志刚[2] 杜喆[2] 孟鑫[1] 

机构地区：[1]河南科技大学机电工程学院,河南洛阳471003 [2]河南科技大学医学技术与工程学院,河南洛阳471003

出　　处：《光学技术》2017年第3期234-238,共5页Optical Technique

基　　金：国家自然科学基金(51305128);河南省科技攻关项目(16212210049);2015HNCXY003

摘　　要：传统的自动对焦梯度函数评价算法在对细胞显微图像进行对焦时,由于对焦精度不佳,图像受噪声的影响较大。针对此问题,通过对不同对焦评价函数性能进行对比,提出了一种基于Sobel梯度函数改进的自动对焦评价算法——Sobel4direction_Brenner梯度函数算法。对对焦评价技术设计指标和算法的通用性进行了验证,结果表明:在调焦时该算法较传统对焦梯度函数能更好的抑制噪声,并且使电机在爬山算法的搜索对焦中具有更小的对焦搜索区间范围,获得的细胞显微图像的清晰度评价值也更高。Due to the poor focusing accuracy, the cell microscopic images are greatly influenced by the noise when they are focused on by traditional auto-focus evaluation algorithm of gradient function. Aiming at the problem, by comparing the performance of different focusing evaluation function, an improved auto-focus evaluating algorithm is proposed based on Sobel gradient function-Sobel4direction_Brenner algorithm of gradient function. After verifying the design inde- xes of focusing evaluation technology and the university of algorithm, the algorithm can restrain the noise better than the traditional one when focusing on the images. The motor has smaller focus search range in the search focus of hill-climbing algorithm. The cell microscopic images are obtained with higher evaluation value of sharpness.

关 键 词：应用光学 自动对焦 对焦精度 显微图像 Sobel梯度函数 爬山算法 Sobel4direction_Brenner梯度函数 

分 类 号：TP273.5[自动化与计算机技术—检测技术与自动化装置]