How Infrakit simplified large-scale point-cloud visualization with LumiDB’s query-driven engine
Infrakit partnered with LumiDB to streamline its point-cloud workflows and replace file-based LOD pipeline. The integration delivered faster rendering, sub-second cross-sections, and a simpler, more stable architecture for handling large LiDAR and drone datasets. This case study outlines how the change was implemented and what improved for engineers and end users.
Written by
Published on
Do not index
Do not index
Infrakit builds a digital construction platform that connects design, field, and machine data across large infrastructure projects. As the company expanded its use of LiDAR and drone scanning to improve project verification, its development team, led by Senior Software Developer Kim Toivonen, faced growing challenges in processing and rendering the enormous point-cloud datasets behind the platform. Each project generated tens of gigabytes of spatial data in varying formats, forcing engineers to rebuild ingestion and level-of-detail (LOD) logic for every case.
The challenge
Custom infrastructure couldn’t scale with growing data volumes
The internal setup was functional but brittle. LOD generation required constant tuning, preprocessing was slow, and browser-based rendering struggled to keep up.
As projects grew in size and complexity, it became clear that maintaining their own 3D data pipeline was unsustainable. The team began looking for a dedicated solution that could handle scale and simplify the workflow without reducing control.
The solution
LumiDB introduced a database designed for massive point clouds
Instead of expanding internal tools, Infrakit integrated LumiDB’s query-based point-cloud engine to replace file-level operations entirely. LumiDB ingests multiple LAS, LAZ, or E57 files and consolidates them into a single spatial database, allowing the system to query across sources as one unified dataset.
This change removed redundant preprocessing, eliminated file management across the stack, and allowed developers to request only the specific regions or resolutions needed for visualization.
The results
Implementation completed in three days, eliminating thousands of lines of code
Ingestion connected easily: LumiDB’s upload API required only a single HTTP call from Infrakit’s existing preprocessing script. On the query side, LumiDB’s database model replaced custom endpoints for level-of-detail retrieval and filtering.
With that change, Infrakit retired thousands of lines of code. Backend and frontend systems became leaner and easier to maintain. Rendering stability improved instantly.
After implementation, large point clouds loaded directly in browsers without stutter or delay. Cross-sections from multiple data sources rendered in under one second, and visual performance remained consistent even for large infrastructure datasets
Moving toward time-series visualization and multi-sensor analytics
With LumiDB providing a scalable 3D data backbone, Infrakit is expanding its visualization tools to combine point clouds across time and integrate machine telemetry from construction equipment.
Ready to simplify your 3D visualization stack?
See how LumiDB can eliminate complexity and scale point-cloud rendering in your browser-based platform.
Book a demo to explore how it integrates with your existing data pipeline.
Written by
