Urban Planning and AEC

Oslo (Norway) – Multi-Source Alignment for BIM Creation

•Task: Integrate diverse data sources to deliver unmatched accuracy and scalability in BIM modeling within 36 hours.

•Combined Google Earth, NavVis LiDAR scans, 360-degree panoramic images, and OpenStreetMap (OSM) data to create a consistent, detailed point cloud. Semantic alignment ensured high-precision LOD 200-300 BIM outputs for urban planning and AEC applications

•Urban development, renovation planning, and scalable BIM solutions for large-scale projects

Problem Statement:

Legacy mapping approaches suffer from inaccurate data and limited detail, while manual processing of point clouds remains costly and inefficient

This creates a need for advanced technologies that can seamlessly integrate and validate diverse geospatial data sources

Data Integration:

1. Google Earth (validation use only)

Satellite images/mesh provide consistent geometric baseline (0.3m)

2. Point-cloud from NavVis scanner provided by a client

NavVis scan provides LiDAR based point-cloud (0.03m, low density)

3. Dense point-cloud from panoramic images only

Panoramic images based point-cloud (0.3m, high density, wide coverage)

4. Dense point-cloud from Openstreetmap (OSM)

OSM provide consistent semantic baseline (0.3m)

Data Fusion and Alignment:

Aligned dense point cloud results in the following:

•Consistent geometry (verified with Google Earth)

•Semantic/geometric priors based on Openstreetmap (road, buildings, etc.) extended areas not captured by a scan

•Accurate geometry based on LiDAR for indoor/outdoor for main area of interest, baseline LOD (200)

•10x level of details based on panoramic-image based point-cloud for higher LOD (300) recognition/modeling

Data Processing Pipeline:

•2D object segmentation indoor (raw)

•At the moment, we are working on improving the technology for recognizing any object based on the data provided

Data Processing Pipeline:

3D building level BIM object segmentation (raw)

Export:

•Segmented point cloud data automatically converted into Revit-compatible BIM layers.

•Layers include detailed 3D models for indoor and outdoor structures, roads, and extended areas from OSM.

•Revit families with LOD 200-300 for immediate use in renovation or urban planning.

Summary:

The integrated mapping solution delivers BIM models within 36 hours, combining 0.03m-precise LiDAR data with dense panoramic point clouds to achieve LOD 300 detail – offering a scalable approach that maintains both geometric accuracy and semantic richness while significantly reducing traditional mapping time and costs

For more information on the technologies used, please follow the link

Need precise and scalable AI solutions?

Urban Planning and AEC

Oslo (Norway) – Multi-Source Alignment for BIM Creation

Problem Statement:

Legacy mapping approaches suffer from inaccurate data and limited detail, while manual processing of point clouds remains costly and inefficient This creates a need for advanced technologies that can seamlessly integrate and validate diverse geospatial data sources

Data Integration:

2. Point-cloud from NavVis scanner provided by a client

3. Dense point-cloud from panoramic images only

Panoramic images based point-cloud (0.3m, high density, wide coverage)

4. Dense point-cloud from Openstreetmap (OSM)

OSM provide consistent semantic baseline (0.3m)

Data Fusion and Alignment:

Aligned dense point cloud results in the following:

•Consistent geometry (verified with Google Earth)

•Semantic/geometric priors based on Openstreetmap (road, buildings, etc.) extended areas not captured by a scan

•Accurate geometry based on LiDAR for indoor/outdoor for main area of interest, baseline LOD (200)

•10x level of details based on panoramic-image based point-cloud for higher LOD (300) recognition/modeling

Data Processing Pipeline:

•2D object segmentation indoor (raw)•At the moment, we are working on improving the technology for recognizing any object based on the data provided

Data Processing Pipeline:

3D building level BIM object segmentation (raw)

Export:

•Segmented point cloud data automatically converted into Revit-compatible BIM layers.

•Layers include detailed 3D models for indoor and outdoor structures, roads, and extended areas from OSM.

•Revit families with LOD 200-300 for immediate use in renovation or urban planning.

Summary:

For more information on the technologies used, please follow the link

Legacy mapping approaches suffer from inaccurate data and limited detail, while manual processing of point clouds remains costly and inefficient

This creates a need for advanced technologies that can seamlessly integrate and validate diverse geospatial data sources

•2D object segmentation indoor (raw)

•At the moment, we are working on improving the technology for recognizing any object based on the data provided