An Accurate Surface Measurement Method Based on Depth-from-focus Considering Magnification Variations

Errors resulting from magnification variations of optical system are largely generated in threemensional shape measurements based on depth-from-focus. Focus measures are wrongly or ambiguously tracted, due to the variations. In this paper, a DFF-based methodology considering these magnification riations is proposed to make accurate measurements in surface morphology with high depth discontinuity. The age magnifications are represented as magnification factors, which are computed with the ratio of diagonal age lengths of a rectangular calibration block, according to optical system steps. The proposed image libration with the least square method makes the image size of measured object features identical, in spite of agnification variations of optical system. So, the accurate and precise applications of focus measures are made the calibrated images. The presented method with a linear magnification calibration has been verified through e reliable experiments of step-shaped objects.