聚焦函数性能评价指标设计及最优函数选取

Design of Evaluation Index for Auto-Focusing Function and Optimal Function Selection

显微镜自动聚焦其核心问题在于聚焦函数的设计及选取,针对目前自动聚焦函数选取缺乏客观定量标准的问题,在分析聚焦函数曲线特征的基础上,设计了陡峭区宽度、清晰度比率、陡峭度、平缓区波动量、局部极值因子和灵敏度等6个定量评价指标,这些指标不仅能用于最优函数的选取,还可以为新的聚焦函数的设计提供理论依据。然后,以这些指标为依据,对目前已经提出的12种典型聚焦函数的性能进行了定量评估,评估过程考虑了图像内容多寡对聚焦函数性能的影响。最后以两阶段聚焦过程为例给出了最优函数:在粗聚焦阶段,方差函数为最优函数,在精细聚焦阶段,Brenner函数为最优函数。

Selection of the optimal focusing function from a given set of focusing functions is an important issue in image-based auto-focusinging that is usually used in automatic image acquisition systems,such as automatic microscopes,vision-based micro-operating robots,etc.However,there is no quantitative index available for evaluating the performance of the focusing functions.Six quantitative evaluation index,including the width of steep part of focusing curve,the ratio of sharpness,the steepness,the variance of flat part of focusing curve,the factor of local extreme and the sensitivity are designed.These index can not only been used in the selection of optimal focusing function,but also provide theoretical basis for the design of new focusing function.Then,twelve typical focusing functions are evaluated based on the quantitative evaluation index.Results show that the variance function is the optimal focusing function in coarse auto-focusing,while the Brenner function is the optimal focusing function in fine auto-focusing.