Loading
Research · Thread 05
Spatial asset intelligence
Turning one handheld building scan into a per-object, queryable asset registry.
A single walkthrough scan should do more than render a 3D model. It should locate every fire extinguisher, smoke detector and AC unit as a clickable, exportable asset - the idea behind MSIP (Municipal Spatial Intelligence Platform), pitched as a fire-safety register built straight from a building scan. The hard part is not detection, it is the camera trajectory: the device exports only ~23 seconds of poses out of a six-minute scan, so the rest is re-derived from raw LiDAR and IMU.
Finding
Re-deriving the camera path
The blocker turned out to be trajectory rather than recognition. Only a fraction of a walk's camera poses ship with the scan, so the rest of the path has to be re-derived from the raw sensor stream before any object can be pinned to a place in the building.
Detection was never the bottleneck. Recovering the full camera path is.
Paper to shipped
Reference
KISS-ICP: In Defense of Point-to-Point ICP (Vizzo, Guadagnino, Mersch, Wiesmann, Behley, Stachniss, IEEE RA-L 2023)Point-to-point ICP against a sparse voxel map, with no IMU and no learned components. The dependency-light baseline for recovering a sensor trajectory from raw LiDAR sweeps.
View abstract
The system performs LiDAR odometry via point-to-point ICP combined with adaptive thresholding for correspondence matching, a robust kernel, motion compensation from a constant-velocity model, and point subsampling, working across sensors and mounting positions at sensor frame-rate without dataset-specific parameter tuning. The authors argue much of the complexity in modern odometry pipelines is unnecessary and a carefully engineered classic approach stays competitive.
Reference
FAST-LIO2: Fast Direct LiDAR-Inertial Odometry (Xu, Cai, He, Lin, Zhang, IEEE T-RO 2022)Tightly-coupled LiDAR + IMU via an iterated Kalman filter on an incremental kd-tree. The robustness step for the fast-rotation and featureless moments where LiDAR-only odometry drops.
View abstract
FAST-LIO2 is a direct LiDAR-inertial odometry framework that registers raw points to the map without feature extraction, tightly fusing inertial measurements through an iterated extended Kalman filter, and maintains the map with an incremental k-d tree (ikd-Tree) supporting incremental insertion and dynamic re-balancing. It demonstrates high accuracy and robustness at high odometry rates across structured and unstructured environments and varied LiDARs.
In progress
Trajectory recovery on the Onyx Tower reference scanLiDAR-only ICP and tight-coupled LiDAR-inertial odometry both drift on the full walk: FAST-LIO2 holds for part of it, then diverges at a fast rotation. What works is frame-to-map localization against the device's own metric reconstruction, which gives drift-free per-frame poses. Wiring those poses into the mask-projection stage is next, and the payoff is lifting per-frame object masks to per-object 3D across the whole scan, rather than the small share of assets anchored by direct camera evidence today.
Proposed
MSIP, a Municipal Spatial Intelligence Platform pilotThe productised pitch: one walkthrough scan in, a queryable fire-safety register out - compliance PDF, IFC, COBie, GeoJSON. On the Onyx Tower reference scan, an automated QA gate confirms fire and life-safety assets across several classes, and the compliance headline reports only the operator-verified ones, keeping colour-detected candidates separate until a person confirms them. Sent, discussed, and PoC-approved by the Municipality client before BARNES's own stakeholders paused the pilot.
Proposed
A living digital twin for a US coastal cityThe same reality-first premise, scaled from one building to a whole city, proposed for a compact coastal city in Florida. MSIP reads a fire-safety register off the scan; this reads a city off it. Planning, resilience and public consultation all draw on one photoreal model. Flood exposure is read at street and property level, and evacuation routes are tested in advance of the event they are for. Proposed developments drop into the real city, so height, shadow and street-level impact can be reviewed before approval, and a resident can open a public explorer and see what is coming near them. A small city is the right first target for the same reason the Onyx Tower was, it is compact enough to finish, and a finished pilot is what convinces anyone. Proposal only, nothing built.