Automated design of freeform imaging systems

TLDR

Automated design of imaging systems with minimal human effort has long been a goal, yet selecting an appropriate starting point remains challenging and differs from the traditional surface‑coefficient optimization approach. The authors propose a novel point‑by‑point design framework that automatically generates high‑performance freeform optical systems. The framework requires only a set of planes as input, automates the selection of starting points and optimization, and was demonstrated by designing two high‑performance systems. The automated process markedly reduces human effort, eliminates the need for advanced design expertise, and enables the creation of freeform systems for remote sensing, telescopy, microscopy, spectroscopy, virtual reality, and augmented reality.

Abstract

The automated design of imaging systems involving no or minimal human effort has always been the expectation of scientists, researchers and optical engineers. In addition, it is challenging to choose an appropriate starting point for an optical system design. In this paper, we present a novel design framework based on a point-by-point design process that can automatically obtain high-performance freeform systems. This framework only requires a combination of planes as the input based on the configuration requirements or the prior knowledge of designers. This point-by-point design framework is different from the decades-long tradition of optimizing surface coefficients. Compared with the traditional design method, whereby the selection of the starting point and the optimization process are independent of each other and require extensive amount of human effort, there are no obvious differences between these two processes in our design framework, and the entire design process is mostly automated. This automated design process significantly reduces the amount of human effort required and does not rely on advanced design skills and experience. To demonstrate the feasibility of the proposed design framework, we successfully designed two high-performance systems as examples. This point-by-point design framework opens up new possibilities for automated optical design and can be used to develop automated optical design in the areas of remote sensing, telescopy, microscopy, spectroscopy, virtual reality and augmented reality.