A new bi-modal dataset called ShotcreteDepth has been released, combining stereo RGB and LiDAR data for robotic depth perception in construction environments. This dataset is designed to address challenges posed by harsh conditions such as high turbidity and poor illumination, which often lead to incomplete and noisy sensor measurements. ShotcreteDepth includes over 11,000 synchronized data samples, with 220 annotated for evaluation, and is accompanied by a lightweight annotation tool for LiDAR point clouds, supporting research in depth estimation and matching. AI
IMPACT Provides a specialized dataset for improving robotic depth perception in challenging construction environments.
RANK_REASON The item describes a new dataset released for research purposes, including a paper and associated code. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Hugging Face Daily Papers →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
