Multi-Modal Non-Line-of-Sight Passive Imaging.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Signal Processing

Background:

• Traditional imaging relies on line-of-sight propagation, failing in non-line-of-sight (NLOS) scenarios.
• Diffusive surfaces scatter light, rendering conventional lens-based imaging ineffective for obscured objects.
• Partially coherent light fields are common in real-world environments, posing challenges for imaging.

Purpose of the Study:

• To develop a passive imaging method for reconstructing 2D projections of objects in NLOS conditions.
• To exploit the 4D spatial coherence function of scattered light for object reconstruction.
• To create a framework for fusing multi-modal information for scene reconstruction in challenging NLOS regimes.

Main Methods:

• Utilizing the 4D spatial coherence function of light reflected from a diffusive wall.
• Formulating a multi-criteria convex optimization problem for reconstruction.
• Leveraging established optics models for light propagation and scattering, incorporating image sparsity.
• Developing an algorithm based on the alternating direction method of multipliers (ADMM).

Main Results:

• Successful reconstruction of 2D projections of obscured objects using reflected light.
• Demonstration of a passive approach compatible with partially coherent fields.
• A convex optimization framework effectively fuses intensity and spatial coherence information.
• An efficient ADMM-based algorithm for solving the reconstruction problem.

Conclusions:

• The proposed method advances passive imaging capabilities in challenging NLOS environments.
• The framework enables multi-modal information fusion for robust scene reconstruction.
• This technique is promising for applications where direct line-of-sight is obstructed.