InfraSpect
InfraSpect is a modular ROS 2 system that ingests recorded or live LiDAR data, reconstructs a 3-D surface model, extracts geometric features, and tracks changes over time. It includes a browser dashboard and an interactive 3-D viewer.
Pipeline
Ingestion → preprocess → registration → reconstruction
Outputs
Point clouds, meshes, features, snapshots, change reports
UI
Live dashboard + in-browser 3-D twin viewer


Key capabilities
A practical end-to-end pipeline for 3-D monitoring and analysis.
- Bag replay into standard PointCloud2 streams (ROS 1/2)
- Noise removal + voxel downsampling + ground extraction
- Scan alignment with ICP (optional coarse stage)
- Surface reconstruction and versioned snapshots (cloud/mesh/JSON)
- Geometric features: bounding boxes, area/volume, roughness, density
- Temporal change detection with severity classification + RViz markers
Architecture
Modular ROS 2 pipeline: each module is an ament package with its own config and launch.
Rosbag / Sensor
│
▼
┌─────────────┐ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐
│ Ingestion │───▶│ Preprocess │───▶│ Registration │───▶│ Twin Build │
└─────────────┘ └──────────────┘ └──────────────┘ └──────┬───────┘
│
┌─────────────────────────┘
▼
┌──────────────┐ ┌──────────────┐
│ Features │───▶│ Storage │
└──────────────┘ └──────┬───────┘
│
┌────────────────────┘
▼
┌──────────────┐ ┌──────────────┐
│ Change Det. │───▶│ Viz / RViz │
└──────────────┘ └──────────────┘
┌──────────────────────────────────┐
│ Dashboard + 3-D Viewer │
└──────────────────────────────────┘Modules
Pipeline stages mapped to responsibilities (as organized in the repository).
- 1) Ingestion — bag playback, validation, TF, diagnostics
- 2) Preprocessing — outlier removal, voxel grid, ground plane
- 3) Registration — ICP alignment, cumulative transforms
- 4) Twin construction — accumulation + Poisson mesh + snapshots
- 5) Feature engineering — geometry features and summaries
- 6) State storage — versioned catalogue with provenance
- 7) Change detection — displacement analysis and severity
- 8) Visualization — RViz markers + formatted reports
- Dashboard — Flask UI + Three.js 3-D viewer
Quick start (repo)
InfraSpect includes a one-command launcher; the repo README has full setup instructions.
The pipeline is built around ROS 2 (Jazzy/Humble), Python, and Open3D. Once running, the dashboard is served locally and includes an interactive 3-D viewer.