This Letter discusses the generation of 3D-printed micro-optics to obtain the desired beam profile from a multimode vertical-cavity surface-emitting laser (VCSEL) with a significantly reduced divergence angle via the usage of high-resolution two-photon polymerization. Due to the low cost and compact packaging, the VCSEL array is a novel light source for structured-light projection. Particularly for long-distance 3D sensing applications, a greatly reduced divergence angle ensures that a good signal with a sufficiently large number of photons can be recorded, and the projected illumination spots do not overlap. Therefore, exact laser beam characterization and appropriate physical modeling are required in accurate production of an optimal collimator lens. Furthermore, elliptical beam profiles with different orientations can solve the correspondence problem and improve the post-processing speed and robustness in structured light. To generate this special type of beam profile and verify the optical design process, this Letter describes thoroughly the optical prototyping process starting from the beam characterization, the optical design to the production of the two-photon polymerized optics, and its validation. The test of the beam profile and divergence confirm a good match of the produced optics with the physical optical simulation in Zemax. The collimator transforms the input laser beam divergence angle of 324 mrad to an output angle of 20 mrad only.

This paper addresses the effect and influence of wave optical aberrations and surface roughness on the chromatic confocal signal and resulting measurement errors. Two possible approaches exist for implementing chromatic confocal imaging based on either refraction or diffraction. Both concepts are compared and an expression for the expected chromatic longitudinal aberrations when using a diffractive optical element is derived. Since most chromatic confocal sensors are point sensors, the discussion on wave-optical aberrations is focused on spherical aberrations. Against common belief, the effect of spherical aberrations cannot be eliminated in the calibration process using for instance a piezo mounted mirror. It will be shown in the following that even a diffraction limited system with peak to valley spherical aberration smaller than 0.25 wavelength suffers from measurement errors. Experimental results will be shown to highlight this important issue. In order to develop a deeper understanding of the underlying physics, a wave-optical simulation environment has been realized. This wave-optical model furthermore enables the investigation of the influence of roughness. Herethereto the correct choice of numerical aperture when investigating a rough surface is based on a heuristic approach. Using the wave-optical simulations an explanation for the increased noise when employing a low numerical aperture to examine rough surfaces will be derived. Furthermore, a formula is presented to support the selection of the correct numerical aperture with regards to the roughness parameters of the surface under investigation.

This paper presents a wave-optical model for chromatic confocal signals based on multimode optical fibre data acquisition. It is shown that the emitted light from a multimode fibre can be modelled as a point source grid. The confocal signal can be obtained by incoherent summation of the intensities of the reflected fields in the detector plane. In this manner, this two-dimensional wave-optical model encompasses the temporal incoherence of the light source due to the usage of a broadband light source and the spatial incoherence due to the application of a multimode fibre. In fact, this model enables the modelling of different degrees of coherence. Therefore, to our knowledge, it represents the most thorough modelling tool for chromatic confocal signals. Physical parameters such as the spectral distribution of the polychromatic light source, the lateral distribution of point sources on the grid, and the strength of longitudinal aberrations, and their influences on the chromatic confocal signal have been evaluated in detail. Consideration of all these parameters results in simulated chromatic confocal signals, which almost perfectly match the experimentally recorded chromatic confocal signals. The results obtained will be very useful for the development of a deeper understanding of chromatic confocal sensing, to optimise the setup parameters with regard to a particular application and to reinterpret the results, possibly in combination with artificial intelligence algorithms, in order to increase the measurement precision and to reduce the demands imposed on the manufacturing quality of the optical elements.

No abstract