Virtual Wave Optics for Non-Line-of-Sight Imaging (1810.07535v2)

Abstract: Non-Line-of-Sight (NLOS) imaging allows to observe objects partially or fully occluded from direct view, by analyzing indirect diffuse reflections off a secondary, relay surface. Despite its many potential applications, existing methods lack practical usability due to several shared limitations, including the assumption of single scattering only, lack of occlusions, and Lambertian reflectance. We lift these limitations by transforming the NLOS problem into a virtual Line-Of-Sight (LOS) one. Since imaging information cannot be recovered from the irradiance arriving at the relay surface, we introduce the concept of the phasor field, a mathematical construct representing a fast variation in irradiance. We show that NLOS light transport can be modeled as the propagation of a phasor field wave, which can be solved accurately by the Rayleigh-Sommerfeld diffraction integral. We demonstrate for the first time NLOS reconstruction of complex scenes with strong multiply scattered and ambient light, arbitrary materials, large depth range, and occlusions. Our method handles these challenging cases without explicitly developing a light transport model. By leveraging existing fast algorithms, we outperform existing methods in terms of execution speed, computational complexity, and memory use. We believe that our approach will help unlock the potential of NLOS imaging, and the development of novel applications not restricted to lab conditions. For example, we demonstrate both refocusing and transient NLOS videos of real-world, complex scenes with large depth.

• The paper introduces the phasor field concept to convert the NLOS challenge into a virtual LOS problem through wave-based modeling.
• It employs the Rayleigh-Sommerfeld diffraction integral to efficiently reconstruct complex scenes while overcoming single-scattering limitations.
• The method achieves high-quality 3D reconstructions in challenging environments, paving the way for advanced applications in navigation, security, and medical imaging.

Insights into Virtual Wave Optics for Non-Line-of-Sight Imaging

The paper "Virtual Wave Optics for Non-Line-of-Sight Imaging" presents a transformative approach to Non-Line-of-Sight (NLOS) imaging, which traditionally entails reconstructing images of objects hidden from direct view by analyzing light indirectly reflected off a secondary relay surface. Previous methodologies in NLOS imaging have been hampered by significant constraints including the consideration of only single scattering light, the neglect of occlusions, and the assumption of Lambertian reflectance. This research offers an innovative solution by redefining the NLOS challenge as a virtual Line-Of-Sight (LOS) imaging problem, employing a novel phasor field concept to characterize the variability in irradiance arriving at the relay surface.

Key Contributions and Methodology

The central innovation of this work lies in the introduction of the phasor field, a mathematical abstraction that models the propagation of NLOS light as a wave, which is subsequently solved via the Rayleigh-Sommerfeld diffraction integral. This approach enables the reconstruction of complex scenes characterized by diverse material properties, significant depth variations, and the presence of both multiply scattered and ambient light. Crucially, the authors demonstrate NLOS reconstruction without resorting to explicit light transport modeling, thus enhancing computational efficiency and reducing memory requirements. By integrating fast computational algorithms, the method surpasses existing solutions in execution speed and computational complexity.

Theoretical and Practical Implications

The theoretical significance of this research lies in its departure from conventional NLOS imaging techniques by leveraging wave-based modeling of light propagation. By transforming the problem into a virtual LOS framework, traditional imaging techniques and algorithms become applicable, thus expanding the potential capabilities of NLOS imaging significantly. The practical implications are profound, as this method unlocks new possibilities for applications beyond controlled laboratory settings. For instance, the paper illustrates the generation of both refocused and time-resolved NLOS videos in complex real-world scenarios, potentially enabling advancements in sectors like autonomous navigation, security, and medical imaging, where indirect visualization is crucial.

Results and Comparisons

The authors provide a comprehensive evaluation of their method, highlighting its robustness and reliability under challenging conditions that include strong ambient illumination and multipath interference. In contrast to the limitations of earlier techniques, including narrow field accuracy and high noise susceptibility, the wave-based approach maintains performance integrity across diverse environmental and observational settings. The paper’s results demonstrate 3D reconstructions of scenes with a clarity and depth range previously unachievable, setting a new benchmark for NLOS imaging quality.

Future Directions

This work suggests several avenues for future research. One promising direction involves further enhancement of the phasor field conceptual framework to improve image fidelity, especially in more intricate scenes. Additionally, integrating this NLOS imaging technique with existing LOS technologies could foster hybrid systems capable of unprecedented observation capabilities. Exploring alternative wave-based computational strategies to further decrease processing times or boost resolution could also be fruitful.

In conclusion, the presented method paves the way for significant advancements in NLOS imaging by bridging the gap between indirect and direct imaging methodologies through virtual wave optics. The implications of this research are significant, promising the development of next-generation imaging systems that are both versatile and adaptive to real-world complexities.