Search

Highlights • We present a novel flatness test which only using two flats. • Multi-shift and rotation steps are needed. • A pixel level spatial resolution can be achieved. • Analysis to tilt error and rotation error is presented. • Simulations and experimental absolute results are given. Abstract In classical three-flat test, error in reference surface can not be identified fully. What we obtain is only the absolute radial profile instead. In addition, at least three flats with similarly high accuracy are required. The test procedures are very complex and it takes a long time and requires high precision adjustment. In this paper, we proposed a new method for absolute test of flatness. There are only two flats used in our method and it does not need precise alignment of the specimens. The method requires multiple translations of the test surface and one 90° rotation step. The relative tilt during translation is measured by a compact wavefront interferometer. We analyze the effect of noise, tilt estimation error and rotation error on reconstruction result. The method is experimentally verified and compared with the direct interferometric result using high precision reference surface. [ABSTRACT FROM AUTHOR]

• Measurement of large area micro topography with subaperture stitching. • Detection of edge features with slope threshold and projection on OXY plane. • Fast correction of 2D lateral coordinates by solving a linear least-squares. • Correction of height coordinates by solving a linear least-squares problem. • No need for high precision scanning stages with "sliding" eliminated by stitching. Microstructured surfaces over large areas are typical in diffractive optics and biomimetic surfaces. The topography measurement is challenging due to the multi-scale nature. In this paper, an efficient stitching method is proposed for fusion of multiple subapertures to produce a large area topography. The first step is detecting edge features and then projecting them on the OXY plane. The next step is correction of lateral coordinates by registering the edge curves in overlapping regions. The change of lateral coordinates is linearly related to the lateral shifts and clocking angle which are estimated by simply solving a linear least-squares problem. In the final step, height change is also linearly related to the piston and tip-tilt of subapertures. Hence it is corrected by minimizing the height difference in overlapping regions, which is again modeled as a linear least-squares problem. The stitching method is experimentally verified on a holographic sample. Totally seven by seven subapertures are measured with an interference microscope and seamlessly stitched together. [ABSTRACT FROM AUTHOR]