New LiDAR method reconstructs hidden scenes through transparent occluders

Researchers have developed a new framework to reconstruct 3D scenes from single-photon LiDAR data, even when partially obscured by transparent materials. This method addresses the challenge of mixed returns from both foreground occluders and hidden objects by analyzing the time-resolved histograms of single-photon LiDAR. The proposed state-aware framework infers echo states for each ray to guide a neural field, improving the accuracy of hidden scene depth and point-cloud reconstruction compared to existing methods. A new dataset of paired single-photon LiDAR captures with and without occlusions was also introduced to validate the approach. AI

IMPACT Enables more robust 3D perception in robotics and autonomous systems operating in complex environments with occlusions.

RANK_REASON Academic paper detailing a new method for LiDAR reconstruction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• DagsHub
• Gotit.pub
• Hugging Face
• lidar
• Litmaps
• neural field
• ScienceCast
• Scite
• Single photon lidar signal attenuation under boreal forest conditions

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →