Weak non-line-of-sight target echoes extraction without accumulation

Area of Science:

• Optics and Photonics
• Signal Processing
• Computer Vision

Background:

• Non-line-of-sight (NLOS) technology enables visualization of hidden objects using returned photons, with applications in autonomous driving and rescue operations.
• Current NLOS methods struggle with weak signals due to laser attenuation and multiple reflections, as direct accumulation amplifies noise.
• Effective separation of weak target echoes from noise is crucial for advancing NLOS applications.

Purpose of the Study:

• To develop and demonstrate novel denoising methods for NLOS imaging that overcome the limitations of traditional signal accumulation techniques.
• To improve the detection and tracking capabilities for weak target echoes in challenging NLOS environments.
• To validate the efficacy of proposed methods through experimental demonstration.

Main Methods:

• Exploration of two denoising techniques based on temporal correlation features, avoiding signal accumulation.
• Proposal of a dual-detector method utilizing software operations to enhance weak signal detection.
• Introduction of a pipeline method for NLOS target tracking in sequential histograms.

Main Results:

• Successful experimental demonstration of both proposed denoising methods.
• Effective extraction of the motion trajectory of a hidden object using the developed techniques.
• Validation of improved detection ability for weak signals without simultaneous noise accumulation.

Conclusions:

• The developed denoising methods offer a promising solution for detecting weak signals in NLOS imaging.
• These techniques can significantly improve the performance of NLOS applications by enabling the extraction of subtle target echoes.
• The findings provide a foundation for future practical implementations in fields requiring hidden object detection and tracking.