Revealing hidden scenes by photon-efficient occlusion-based opportunistic active imaging.

Area of Science:

• Computational Imaging
• Optical Science
• Computer Vision

Background:

• Recovering details of hidden scenes from environmental reflections is challenging due to light scattering.
• Diffuse reflections mix spatial information, hindering effective scene reconstruction.
• Existing methods often rely on complex, ultrafast time-of-flight measurements.

Purpose of the Study:

• To develop a novel computational imaging technique for reconstructing hidden scenes.
• To overcome limitations of diffuse reflections and complex measurement requirements in hidden-scene imaging.
• To demonstrate a photon-efficient method for non-time-resolved hidden scene recovery.

Main Methods:

• Utilizing occluding objects to obstruct probe-light propagation and 'undo' spatial information mixing.
• Employing a computational algorithm based on a forward model of three-bounce light propagation.
• Performing non-time-resolved measurements with single-photon detection.

Main Results:

• Successfully reconstructed hidden-surface reflectivity patterns in a meter-scale environment.
• Demonstrated effective scene recovery despite diffuse reflections.
• Achieved photon-efficient imaging without ultrafast time-of-flight measurements.

Conclusions:

• The developed technique offers a promising new imaging modality for recovering hidden scenes.
• This approach obviates the need for complex, high-speed measurements.
• The method has significant potential implications for various imaging science applications.