3D Tiling Pipeline

3D Buildings Tiler

Use this command line tool to tile your CityGML or KML/COLLADA into 3D Tiles.

As the creators of 3D Tiles, we leverage its full potential at every step in our pipeline to handle millions of textured buildings, trees, and other features. Massive datasets are optimized with varying levels of resolution, geometry and texture batching, and instancing for reduced CPU and GPU overhead. This all saves bandwidth, improves performance, and minimizes memory usage so you can visualize massive cities, even on mobile devices!

The command line interface is easy to integrate into any platform, including Linux, Windows, AWS, and Azure. Convert your 3D building data to an efficient tileset by running a single command. See the full documentation.

High performance

Computationally intensive steps are multithreaded, from the initial parsing to the terrain clamping and tiling, giving the tiler a significant performance advantage. The tiler scales up using all available CPU cores, and out-of-core memory management handles any size dataset, even on commodity hardware.

Pixel-perfect precision

The tiler outputs a precise geometric error so you can be confident about the accuracy at each level of detail. We preserve the highest resolution data at the leaf tiles so nothing is lost. Even the terrain clamping step makes no compromises.


Preserve metadata for runtime analysis

The 3D buildings tiler preserves metadata from CityGML and KML/COLLADA so you can use this information for dynamic run-time styling and analysis in your CesiumJS apps.

Set color, size, and visibility using metadata values or custom expressions combining multiple data points for real-time analysis such as determining optimal solar installations or mapping the growth of cities over time.

Adaptive tiling for efficient streaming

Cities come in all shapes and sizes, with sprawling suburbs and dense neighborhoods. To account for this, the tiler computes non-uniform tiles that consider the geometry and texture distribution in the dataset, yielding a balanced tileset. Tiles can even overlap and fit around buildings instead of splitting them. This results in far better streaming performance than the traditional uniform 2D grid approach.

Reichstag building from CityGML dataset of 540,000 textured Berlin buildings, tiled with the best visual quality at all levels of detail.

Clamp buildings to terrain

Accurately clamping CityGML and KML/COLLADA datasets to the underlying terrain is crucial for applications such as city planning, flood modeling, and line of sight analysis.

The terrain clamping algorithm finds the best fit to avoid artifacts such as floating or sunken buildings, even over steep slopes. You can be confident that what you're seeing corresponds to the ground truth.

The tiler extends building geometry when necessary to fix buildings to terrain.

Opt in for additional optimizations

Draco compression for geometry

Google's Draco compression is 3D topology-aware and reduces geometry sizes by up to 90% without compromising visual quality, performing better than general-purpose algorithms like Gzip.

Decoding in CesiumJS is fast with WebAssembly and decoding in parallel on the CPU and GPU.

Learn more about Draco compression in Cesium

WebP images for textures

WebP images are on average 30% smaller in size than JPEG and PNG images of the same quality—even with transparency.

WebP is the image format designed by Google for the web and is ideal for most modern browsers.

Learn more about WebP in Cesium

Interested in on-premises 3D tiling?

Contact our sales team to request an evaluation, or ask about custom tiling pipelines and services for your 3D data.

Contact us