Three-dimensional integral imaging in photon-starved environments with high-sensitivity image sensors

Area of Science:

• Optics and Photonics
• Image Processing
• Low-light Imaging

Background:

• Imaging in low-light conditions is challenging due to limited photon detection.
• Three-dimensional (3D) integral imaging integrates multiple 2D views, offering potential noise resilience.
• Passive imaging sensors in the visible spectrum struggle with photon scarcity.

Purpose of the Study:

• Investigate the performance of 3D integral imaging under extremely low-light conditions (few to tens of photons per pixel).
• Evaluate and compare imaging capabilities of EM-CCD and sCMOS cameras in this regime.
• Assess the improvement in image quality and visualization for 3D reconstructed images.

Main Methods:

• Experimental verification using EM-CCD and sCMOS cameras.
• Theoretical modeling and fitting for EM-CCD camera parameters.
• Pixelwise calibration for sCMOS camera parameter determination.
• Derivation and experimental corroboration of signal-to-noise ratio (SNR) for both sensors.
• Analysis of contrast-to-noise ratio (CNR) and perception-based image quality estimator (PIQE).

Main Results:

• Successful experimental demonstration of 3D integral imaging with average photon counts as low as a few per pixel.
• Theoretical models for EM-CCD and calibration for sCMOS accurately determined camera parameters.
• Experimental findings corroborated theoretical SNR derivations for both camera types.
• Demonstrated improvement in image quality metrics (CNR, PIQE) for 3D reconstructed images compared to 2D.
• Achieved improved scene visualization, including occlusion removal, in low-light environments.

Conclusions:

• 3D integral imaging is effective for low-light-level conditions, significantly enhancing image quality.
• This work represents the first experimental report on few-photon-level 3D integral imaging for improved visualization.
• The findings provide a foundation for advanced imaging applications in photon-starved scenarios.