Abstract
Imaging through opaque scattering media is critically important in applications ranging from biological and astronomical imaging to metrology and security. While the random process of scattering in turbid media produces scattered light that appears uninformative to the human eye, a wealth of information is contained in the signal and can be recovered using computational post-processing techniques. Recent studies have shown that statistical correlations present in the scattered light, known as ‘memory effects’, allow for diffraction-limited imaging through opaque media without detailed knowledge of (or access to) the source or scatterer. However, previous methods require that the object and/or scatterer be static during the measurement. We overcome this limitation by combining traditional memory effect imaging with coded-aperture-based computational imaging techniques, which enables us to realize for the first time single-shot video of arbitrary dynamic scenes through dynamic, opaque media. This has important implications for a wide range of real-world imaging scenarios.
Highlights
Conventional optical imaging techniques create a one-to-one mapping between object and image planes
We image the coded speckle onto the detector plane. This single, low-contrast coded speckle image I(x,y) represents the superposition of the coded speckle patterns reaching the detector over the course of the acquisition time, and can be described as the Hadamard product of the speckle and the coded aperture pattern
The technique does not require this measurement duration or SNR, though, and we find that the performance remains fairly uniform down to camera frame rates of approximately 1 Hz, at which point the image quality begins to decrease gracefully
Summary
Conventional optical imaging techniques create a one-to-one mapping between object and image planes. A variety of techniques have been developed to allow imaging through opaque materials by filtering out the scattered light[1,2,3], performing wavefront shaping of the light incident on the scatterer[4,5,6], conducting detailed statistical modeling of the scatterer[7], or exploiting intrinsic correlations in the scattered light[8,9,10,11] Of these approaches, only the latter method, known as ‘memory effect’ (ME) imaging[12,13], allows for imaging through highly scattering media without the need for detailed knowledge of or access to the scatterer, object, or illumination. We use a dictionary learning approach[19] to recover multiple high-speed speckle frames from a single acquisition, and independently process these de-multiplexed speckle images to estimate the scene at each frame, yielding an effective framerate that is faster than the detector In this way, we realize single-shot video through an opaque scatterer. As a result of this mathematical similarity, variants of the physical and algorithmic tools developed in the fields of computational and compressive imaging can be brought to bear to code the individual channels and demultiplex them post detection
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
