Researchers Enable Non-Line-of-Sight Imaging Using Consumer LiDAR

Researchers have developed a method to perform non-line-of-sight (NLOS) imaging using consumer-grade LiDAR sensors, enabling the visualization of objects hidden from the sensor's direct field of view. The technique, detailed in a paper posted to arXiv, utilizes a process called motion-induced sampling to overcome the hardware limitations typically associated with consumer electronics.

The Mechanism of NLOS Imaging LiDAR sensors operate by capturing the time-of-flight of light at picosecond resolution. While this capability has previously allowed research-grade LiDAR systems to capture information about hidden objects, consumer-grade sensors have lacked the necessary precision and power to achieve similar results.

To bridge this gap, the research team introduced a multi-frame fusion strategy. This approach is inspired by synthetic aperture radar and burst photography, allowing the system to collect and combine multiple data points over time to reconstruct a scene that is not directly visible.

Motion-Induced Aperture Sampling Central to the breakthrough is the introduction of the motion-induced aperture sampling model. This computational framework is designed to unify and account for three primary variables: the shape of the hidden object, the motion of that object, and the movement of the camera itself.

By interpreting variations in measurements as a function of these dynamic factors, the model can more robustly reconstruct hidden scenes.