The localisation accuracy of axial peaks is an important factor for height determination in a confocal microscope. Several algorithms have been proposed for height extraction in surface topography measurements. However, some algorithms ignore the influence of error and discrete sampling on the accuracy. This paper analyzes the localisation accuracy of some common algorithms under different aberrations and random errors, and discusses the effect of axial scanning interval on the accuracy of each algorithm. Finally, we get the application scope of each algorithm. Our results offer a reference for selecting algorithms for confocal metrology.
We have proposed the component parameters measuring method based on the differential confocal focusing theory. In
order to improve the positioning precision of the laser differential confocal component parameters measurement system
(LDDCPMS), the paper provides a data processing method based on tracking light spot. To reduce the error caused by
the light point moving in collecting the axial intensity signal, the image centroiding algorithm is used to find and track
the center of Airy disk of the images collected by the laser differential confocal system. For weakening the influence of
higher harmonic noises during the measurement, Gaussian filter is used to process the axial intensity signal. Ultimately
the zero point corresponding to the focus of the objective in a differential confocal system is achieved by linear fitting
for the differential confocal axial intensity data. Preliminary experiments indicate that the method based on tracking light
spot can accurately collect the axial intensity response signal of the virtual pinhole, and improve the anti-interference
ability of system. Thus it improves the system positioning accuracy.