Extended depth of focus for high-end machine vision lenses by annular achromatic add-on diffractive elements. (March 2023)

• Record Type: Journal Article
• Authors: Suszek, Jaroslaw; Makowski, Michal; Kolodziejczyk, Andrzej; Wlodarczyk, Filip; Sobczyk, Artur; Nurczyk, Piotr; Duda, Przemyslaw; Starobrat, Joanna; Beck, Romuald
• Is Part Of: Optics and lasers in engineering. Volume 162(2023)
• Publication Date: 2023-03
• Subjects: Imaging -- Depth of focus -- Diffraction -- Optical character recognition; Lasers in engineering -- Periodicals; Optical measurements -- Periodicals; Optics -- Periodicals; Lasers en ingénierie -- Périodiques; Mesures optiques -- Périodiques; Optique -- Périodiques; 621.36605
• DOI: 10.1016/j.optlaseng.2022.107445
• Languages: English
• ISSNs: 0143-8166

Abstract

Highlights: Extended depth of focus of imaging with high-end lenses equipped with static, transparent phase masks clipped on the objective. Achromatic use of sparse diffractive optical element for broadband imaging. Hollow-center, annular design of the phase filters allows unobstructed native quality of the lenses. Easy experimental assessment of imaging quality by measurement of readability of alphanumerical symbols with a neural-network based algorithm. Abstract: In modern machine vision systems high-end lens objectives allow superior imaging on state-of-the-art sensor arrays due to highly corrected optical aberrations and large apertures. The resultant shallow depth of focus is a disadvantage, usually leading to programmatic attempts of extending it, with compromised image resolution. Here we present a simple optical system equipped with novel, annular peripheral add-on diffractive elements, allowing highly extended range of sharp imaging without any pre- or post-processing. The separation of main refractive focusing power from the proposed filters enables sparse kinoform zones leading to easy manufacturing and negligibly low chromaticity in polychromatic imaging. The undisturbed central region of the refractive objective sustains the full native imaging quality of the test Zeiss Otus lens. Convergent numerical and experimental results are presented, showing improved readability of alphanumerical symbols in a wide range of w20 defocus parameters up to 24 wavelengths. Robust neural network-based method of image quality assessment is also proposed based on statistical optical character recognition, allowing high dynamics and easy implementation in experimental use as compared to established methods.