Lidar News Weekly RECAP – August 18-21st
Lidar News Weekly RECAP – August 18-21st
Curated and written by Adam Clark. Adam has spent the past 13 years exploring the world from above by using drones, satellites, and mapping tools to better understand our landscapes.
A Fresh Look at Decommissioning & Offshore Wind Projects Through 3D Vision
Voyis’ Discovery Stereo Vision System empowers ROVs with genuine 3D perception by delivering real-time depth maps and color imagery. Two major recent enhancements elevate this platform: integration of VSLAM (Visual Simultaneous Localization and Mapping) for enhanced spatial awareness without relying on external positioning, and the Deep Vision Optics, a redesigned underwater lens assembly that significantly improves image sharpness, edge clarity, and measurement accuracy, even in low-light or turbid conditions. Together, these advances enable precise, sub-millimeter 3D modeling and live data feedback, optimizing subsea inspection, decommissioning, and offshore wind asset management.
(Image credit: DeepOcean/Voyis)
Automated 3D room mapping and real-time corner detection for Android-based AR systems
This research introduces a novel augmented reality (AR) framework enabling automated 3D room mapping and real-time corner detection directly on standard Android devices. The system leverages AR techniques to dynamically generate spatial layouts with minimal user input, eliminating the need for external hardware or manual processes. It streamlines room modeling by recognizing structural features like corners on-the-fly, supporting real-time construction of spatial geometry. Optimized for mobile execution, it holds promise for applications in interior design, architecture, and indoor navigation using ubiquitous Android platforms.
https://www.sciencedirect.com/science/article/pii/S2590123025024715
Coastal Environments: LiDAR Mapping of Copper Tailings Impacts, Particle Retention of Copper, Leaching, and Toxicity
This study deploys advanced remote sensing technologies to assess historical copper tailings along Michigan’s Keweenaw Peninsula. Using aerial LiDAR bathymetry, multispectral surveys, and UAS (drone) overflights, researchers mapped underwater and coastal sediment dispersal and quantified particle movement toward Buffalo Reef. The technology enabled high-resolution geospatial visualization of mine waste behavior. Laboratory and field analyses then correlated remote sensing insights with copper leaching rates, toxicity impacts on benthic organisms, and environmental risks. The methodology highlights the power of LiDAR and drone mapping in environmental monitoring and tailings remediation planning.
Semantic-Aware Cross-Modal Transfer for UAV-LiDAR Individual Tree Segmentation
This paper introduces a cross-modal semantic transfer framework that segments individual trees by fusing 2D UAV imagery with 3D LiDAR point clouds. It starts with RGB-based instance segmentation using a novel network called MSFFNet, designed for precise tree canopy detection in complex forest scenes. RGB data is aligned with LiDAR through image–point cloud registration, then semantic labels are transferred to the 3D domain via a semantic probability field. Key innovations include multi-scale feature fusion (MLFF), a feature alignment module (FAM) using deformable convolutions, and channel/spatial feature interaction modules.