From raw point cloud to structured BIM model
The construction and AEC industries generate enormous amounts of 3D scan data — but turning that raw data into something usable has always been a manual, time-consuming process. ODA Scan to BIM is built to change that.
What is ODA Scan to BIM?
ODA Scan to BIM is a toolkit for BIM software development that provides automated conversion of 3D laser scan data into parametric BIM models. It is being developed under the ODA Strategic Interoperability Group program — meaning the roadmap is shaped directly by the companies that build on it.
The SDK is currently available as a beta — ODA is actively seeking partners and early adopters to join the Strategic Interoperability Group and shape its development.
The SDK works with both structured and unstructured point clouds and supports all major input formats: .RCP, .RCS, .LAS, .PTS, .PTX, .XYZ, .OBJ. Output is available as .IFC, .Revit, or custom BIM objects.
How it works
The pipeline takes raw point cloud data through four stages:
1. Polygonal surface reconstruction The SDK rebuilds mesh geometry from scanned point cloud data. Mesh models can be edited to remove noise and interference objects. The reconstructed surface is optimized for virtual design workflows and matches the original geometry from which the scan was taken.
2. AEC object classification Polygonal geometry is classified into architectural elements — walls, floors, roof, openings. The SDK identifies and segments objects automatically, turning unstructured mesh data into semantically meaningful building components.
3. B-Rep conversion Classified objects are converted from polygonal representation into precise boundary representation (B-Rep) geometry — the standard format used by BIM applications for editing and analysis.
4. BIM export The SDK automatically converts classified objects into family groups ready for use in IFC, Revit, and other popular BIM software.
What makes it different
ODA Scan to BIM is an open, vendor-neutral toolkit that gives software developers full control over their Scan to BIM pipeline. Built within the ODA ecosystem, it integrates directly with ODA's broader infrastructure for engineering data.
At the core of the pipeline is ML-based semantic segmentation — a machine learning approach trained on real indoor point cloud datasets to identify and classify building elements with high accuracy.
Roadmap
Available now
Point cloud to mesh conversion — formats: RCP, RCS, PTS, XYZ, LAS
Classification of polygonal objects into AEC elements — walls, floors, openings, roofs
Conversion of polygonal model to B-Rep — automatic mode
Coming in 2026
Identifying main building elements enhancements
Large point clouds support
Automatic room detection
Automatic mesh segmentation improvements — torus surface recognition, manual segmentation editing
GUI-based viewing/debug application enhancements
See it live
The best way to understand where the pipeline stands today is to see it live. On April 21, Aleksandr Fedorov, Visualize & Scan to BIM Team Lead, will walk through the full pipeline — semantic segmentation, surface reconstruction, BIM object recognition, and more. Real engineering, live demo, Q&A included.
🌐 Global — April 21 | 12:00 UTC+02:00
🇺🇸 USA — April 22 | 1:00 PM ET / 10:00 AM PT
🇨🇳 Chinese — May 19 | 14:00 Beijing Time
This is part of ODA DevConnect'26 — a free webinar series where ODA engineers walk through real product updates. No marketing, just engineering.
Register → https://devconnect.opendesign.com
Want to discuss your specific use case? After the webinar, you can book a personal session with Visualize & Scan to BIM Team Lead directly.
Book a call → Calendly
