HY
H. Ye
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
2 records found
1
LiDAR-TreeRisk
A transferable, LiDAR-driven framework for storm-event risk of wind damage to individual urban trees
Urban trees deliver ecosystem services until extreme winds turn them into hazards, and municipal managers need a defensible, decision-relevant risk label for the individual tree during real storm events. The mature mechanistic route to tree-level critical wind speed comes from forestry, but it embeds closed-canopy stand-context priors, and existing computational fluid dynamics (CFD)–mechanistic urban hybrids neither neutralise those priors nor drive both branches from one per-tree geometry, reporting a continuous score rather than an actionable label. We present a transferable, per-tree framework, applied to The Hague under the Storm Poly wind event of 5 July 2023. A single airborne-LiDAR-derived per-tree geometry is carried coherently into both a direction-resolved urban CFD exposure field, in which each crown is an alpha-wrapped porous zone, and a ForestGALES turning-moment-coefficient resistance core whose forest-stand priors are neutralised at the coupling interface; risk is expressed as a four-class, directional-and-gust operational taxonomy. We find that which trees are operationally at risk is governed by directional urban exposure and event loading rather than by median resistance. Exposure explains the large majority of the between-tree risk variance, structural capacity acts as a secondary filter, and the at-risk trees concentrate in the low-rise residential district. Results show that the labels are stable under moderate analyst choices and qualitatively consistent with the reported storm damage at the district scale, providing a transferable route from open, city-scale data to a per-tree risk label.
...
Urban trees deliver ecosystem services until extreme winds turn them into hazards, and municipal managers need a defensible, decision-relevant risk label for the individual tree during real storm events. The mature mechanistic route to tree-level critical wind speed comes from forestry, but it embeds closed-canopy stand-context priors, and existing computational fluid dynamics (CFD)–mechanistic urban hybrids neither neutralise those priors nor drive both branches from one per-tree geometry, reporting a continuous score rather than an actionable label. We present a transferable, per-tree framework, applied to The Hague under the Storm Poly wind event of 5 July 2023. A single airborne-LiDAR-derived per-tree geometry is carried coherently into both a direction-resolved urban CFD exposure field, in which each crown is an alpha-wrapped porous zone, and a ForestGALES turning-moment-coefficient resistance core whose forest-stand priors are neutralised at the coupling interface; risk is expressed as a four-class, directional-and-gust operational taxonomy. We find that which trees are operationally at risk is governed by directional urban exposure and event loading rather than by median resistance. Exposure explains the large majority of the between-tree risk variance, structural capacity acts as a secondary filter, and the at-risk trees concentrate in the low-rise residential district. Results show that the labels are stable under moderate analyst choices and qualitatively consistent with the reported storm damage at the district scale, providing a transferable route from open, city-scale data to a per-tree risk label.
SpatiaLLM
Bridging the Gap Between Natural Language and 3D Scans
Recent advances in large language models (LLMs) have expanded natural language reasoning and multimodal understanding but remain limited in grounding with 3D spatial environments. This project addresses that gap by developing a system that enables natural language interaction with indoor spatial data derived from light detection and ranging (LiDAR) point clouds and panoramic imagery provided by the client: ScanPlan. The system processes spatial data through a pipeline that includes room segmentation, geometric analysis, and object clustering. A structured query language lite (SQLite) database stores the structured information, which an AI agent queries using a reasoning framework that translates natural language into actionable commands. The system supports multimodal input, allowing users to interact via text or by selecting objects in 2D panoramas, which are then mapped to 3D point clouds using segment anything model 2 (SAM2). The interface combines a chat function with 2D and 3D viewers, making spatial data accessible to non-experts. While the prototype successfully answers a range of spatial and semantic queries, challenges remain in scaling room segmentation and handling complex multi-room relationships. The project demonstrates a step towards making rich 3Dbuilding data queryable through intuitive, language-based interaction.
...
Recent advances in large language models (LLMs) have expanded natural language reasoning and multimodal understanding but remain limited in grounding with 3D spatial environments. This project addresses that gap by developing a system that enables natural language interaction with indoor spatial data derived from light detection and ranging (LiDAR) point clouds and panoramic imagery provided by the client: ScanPlan. The system processes spatial data through a pipeline that includes room segmentation, geometric analysis, and object clustering. A structured query language lite (SQLite) database stores the structured information, which an AI agent queries using a reasoning framework that translates natural language into actionable commands. The system supports multimodal input, allowing users to interact via text or by selecting objects in 2D panoramas, which are then mapped to 3D point clouds using segment anything model 2 (SAM2). The interface combines a chat function with 2D and 3D viewers, making spatial data accessible to non-experts. While the prototype successfully answers a range of spatial and semantic queries, challenges remain in scaling room segmentation and handling complex multi-room relationships. The project demonstrates a step towards making rich 3Dbuilding data queryable through intuitive, language-based interaction.