Learned Large Field-of-View Imaging With Thin-Plate Optics
Large Field-of-View Imaging With Thin-Plate Optics. We design a lens with compact form factor using one (or two) optimized refractive surfaces on a thin substrate (left). This optimization results in a dual-mixture point spread function (center-left insets), which is nearly invariant to the incident angle, exhibiting a high-intensity peak and a large, almost constant, tail. We show the sensor measurement (center) and image reconstruction (right) in natural lighting conditions, which demonstrate that the proposed deep image recovery effectively removes aberrations and haze resulting from the proposed thin-plate optics. Our prototype single element lens achieves a large field-of-view of 53◦ with a clear aperture of f /1.8 and effective aperture of f /5.4, see paper.
Typical camera optics consist of a system of individual elements that are designed to compensate for the aberrations of a single lens. Recent computational cameras shift some of this correction task from the optics to post-capture processing, reducing the imaging optics to only a few optical elements. However, these systems only achieve reasonable image quality by limiting the field of view (FOV) to a few degrees -- effectively ignoring severe off-axis aberrations with blur sizes of multiple hundred pixels. In this paper, we propose a lens design and learned reconstruction architecture that lift this limitation and provide an order of magnitude increase in field of view using only a single thin-plate lens element. Specifically, we design a lens to produce spatially shift-invariant point spread functions, over the full FOV, that are tailored to the proposed reconstruction architecture. We achieve this with a mixture PSF, consisting of a peak and and a low-pass component, which provides residual contrast instead of a small spot size as in traditional lens designs. To perform the reconstruction, we train a deep network on captured data from a display lab setup, eliminating the need for manual acquisition of training data in the field. We assess the proposed method in simulation and experimentally with a prototype camera system. We compare our system against existing single-element designs, including an aspherical lens and a pinhole, and we compare against a complex multielement lens, validating high-quality large field-of-view (i.e. 53-deg) imaging performance using only a single thin-plate element.