Long sequence single-exposure videography using spatially modulated illumination

Simon Ek,Elias Kristensson,Vassily Kornienko

doi:10.1038/s41598-020-75603-7

Abstract

Frequency recognition algorithm for multiple exposures (FRAME) is a single-exposure imaging technique that can be used for ultrafast videography, achieved through rapid illumination with spatially modulated laser pulses. To date, both the limit in sequence length as well as the relation between sequence length and image quality are unknown for FRAME imaging. Investigating these questions requires a flexible optical arrangement that has the capability of reaching significantly longer image sequences than currently available solutions. In this paper we present a new type of FRAME setup that fulfills this criteria. The setup relies only on (i) a diffractive optical element, (ii) an imaging lens and (iii) a digital micromirror device to generate a modulated pulse train with sequence lengths ranging from 2 to 1024 image frames. To the best of the authors’ knowledge, this is the highest number of temporally resolved frames imaged in a single-exposure.

Highlights

Frequency recognition algorithm for multiple exposures (FRAME) is a single-exposure imaging technique that can be used for ultrafast videography, achieved through rapid illumination with spatially modulated laser pulses
We have demonstrated the compatibility of illumination-based FRAME with long sequence length consisting of up to 1024 individual images: a 250-fold increase from previous demonstrations
This was made possible by greatly reducing the number of essential optical components in the setup to only 3; a diffractive optical element (DOE), an imaging lens and a digital micromirror device (DMD)

Summary

Introduction

Frequency recognition algorithm for multiple exposures (FRAME) is a single-exposure imaging technique that can be used for ultrafast videography, achieved through rapid illumination with spatially modulated laser pulses. The decrease in the average image quality of the sequence in FRAME is a consequence of having to use fewer Fourier components per reconstructed frame. The beams in the two categories can in total be combined into 1984/2 + 32 = 1024 pairs that yield unique interference patterns, setting the maximum sequence length of the setup.

Results

Conclusion