椭偏光学显微成像系统聚焦评价函数的研究

自动聚焦技术在椭偏光学显微成像系统中是一项关键的技术。利用几种常用聚焦评价函数作为研究手段研究了这一过程。实验中采用硅基底表面人血纤维蛋白膜层和多元蛋白质芯片样品作为研究对象，研究了几种函数在自动聚焦过程中函数值的变化，以及噪音、样品区域选择对聚焦效果的影响。实验验证托普拉斯评价函数和Sobel评价函数可以满足系统的要求；采用多幅图像平均的方法可以有效降低噪音．当平均次数达到9次以上．系统噪音的影响就可以被降至有效评价的程度；沿同一锐边，选择不同的调焦区域．对聚焦过程基本没有影响；利用评价函数在实验中确定了倾斜平面成像的最佳聚焦平面，并与理论像平面吻合。实验证明利用聚焦评价函数可以成功实现自动聚焦技术在椭偏光学显微成像系统中的应用。

Study on the Autofocus Evaluation Function in Ellipsometric Imaging System

Auto focusing is key technique in ellipsometric imaging system. The process is studied by use of several autofocus' evaluation functions. FIB on Si wafer and multi-channel proteinchip has been used in experiment. The change of functions value has been studied. The effect of noise and selected area used for auto focusing are evaluated. The experiment shows that Laplacian function and Sobel function are suitable for this system. By using several frames to get average result, the effect of noise has reduced to good level when the number of frame more than 9. Along the same sharp edge, selecting different areas for autofocus, it has no influence on result. Using autofocus evaluation function to get the focus image of tilting plane, and the result agrees with the theoretical plane of image. It has been confirmed that autofocus evaluation function can be used in ellipsometric imaging system.