Cesium for Omniverse v0.25.0

Cesium for Unity Samples v1.18.1

CesiumJS 1.134

CesiumJS

About 25% of the global seafloor has been mapped in high resolution, according to NOAA, leaving a lot of work to be done. It’s not always practical or safe for humans to be involved, and that’s where Terradepth comes in: its autonomous underwater vehicles collect vast amounts of 3D seabed data hosted on Cesium ion Self-Hosted and streamed to its CesiumJS-powered app, Absolute Ocean. 

Terradepth cuts the seafloor-to-browser process from weeks to days. Courtesy Terradepth.

Terradepth Absolute Ocean Platform featuring Cesium Tilers

Traditionally, sending out a crew, collecting data, manually processing it, and making it available to users took several weeks. Using machines for dangerous, repeatable work, processing and sharing data where it’s captured, and employing people for analysis cuts the seafloor-to-browser process to days. 

Terradepth has been using Cesium since Absolute Ocean’s development and release in 2022. Absolute Ocean users collect and manage their data securely so they can recognize geographic physical features, identify human-made objects like sunken wrecks or unexploded ordnance, and observe change over time beneath the waves in immersive 3D. Hydrographers in academia, infrastructure, and national security rely on this growing collection of undersea geospatial data for better, faster decisions, with nearly immediate access to interactive data instead of a static slide or hard drive transfers. 

Skellig Michael, Ireland, in Absolute Ocean. Courtesy Terradepth.

Detail in ocean data varies: 500 m resolution from satellite altimetry, 50-100 m resolution from a ship, and 5 cm resolution from a robot near the seafloor, where human safety is questionable or even impossible. Terradepth’s AUVs use multibeam sonar, magnetometers, and more to collect terabytes of data. More detail means more data, and Terradepth uses Cesium to get it to users efficiently and securely.  

With Cesium ion Self-Hosted, Terradepth stores massive point clouds, GeoTIFFs, and other data types; tiles the point clouds into 3D Tiles; and streams them to Absolute Ocean, which is built on CesiumJS, for sharing with a broad research community or smaller set of decision-makers. Cesium ion Self-Hosted enables Absolute Ocean’s processing and sharing at the edge and supports data residency requirements. In addition to its detailed seabed data, Terradepth accesses Cesium World Terrain and Bing Maps Aerial imagery for accurate geospatial context above sea level. 

“Terradepth collects terabytes of data from the ocean. The sheer size meant customers used to wait weeks or months for their data. But with Absolute Ocean using Cesium Tilers at the edge we can let them see it comparatively instantly—with an intuitive, immersive, and three-dimensional user interface,” said Andrew Lunstad, Terradepth CTO.  

Watch Andrew Lunstad's 2025 Cesium Developer Conference talk to learn more about why Terradepth deploys Cesium ion Self-Hosted.

Developers and the Open Geospatial Ecosystem

Comprehensive, collaborative seafloor data becomes insight into one of the least explored parts of our planet. 

Learn more about hosting, tiling, and streaming massive datasets in your own environment to process at the edge and meet strict security requirements.  

Terradepth Processes Seafloor Data at the Edge with Cesium ion Self-Hosted 

In September, we released Cesium Mars, highlighted MicroStation’s 3D Tiles support and Cesium Native integration, and announced participants in the iTwin Activate cohort focusing on Cesium and 3D Tiles. We also attended the GeoGov Summit and welcomed Khronos members to Philadelphia for the F2F meeting. 

Cesium Mars is a 3D Tileset to enable developers, researchers, and mission planners to create visualizations and simulations of the Red Planet. 

Cesium Mars

We made improvements across Cesium this month. To get notifications about our monthly releases, you can subscribe to the Cesium release roundup thread on our community forum.

CesiumJS 1.134 is now available. Highlights of the release include:

Added Google2DImageryProvider to load imagery from Google Maps. #12913 

Converted voxel raymarching to eye coordinates to fix precision issues in large datasets. #12061 

Fixed vertical misalignment of glyphs in labels with small fonts. #8474 

Fixed flickering artifact in Gaussian splat models caused by incorrect sorting results. #12662 

Sandcastle has been updated at https://sandcastle.cesium.com! The legacy Sandcastle app will remain available through November 3, 2025. 

Cesium for Unreal v2.20.0 is now available. Highlights of this release include:

Added support for rendering glTFs with line primitives. 

Added UCesiumGoogleMapTilesRasterOverlay. 

Added AssetOptions property to UCesiumIonRasterOverlay. 

In addition to the above, this release updates cesium-native from v0.51.0 to v0.52.0. See the changelog for a complete list of changes in cesium-native. 

Check the Cesium for Unreal release and the Cesium for Unreal Samples release for the full list of updates.

Download the Cesium for Unreal plugin from Epic Marketplace

The Google Maps attribution requirements were updated to use the "Google Maps" logo instead of the "Google" logo. Cesium ion and Cesium clients now use the Google Maps logo. If your application includes its own logic for the inclusion of the logo rather than using a logo provided by Cesium, see Google Maps' documentation for guidance about whether your application needs to be changed. 

Support for spherical harmonics for Gaussian splats. 

v1.7.1 is now available, with minor bug fixes. 

Fixed an issue that could cause incorrect materials to be mapped to meshes when loading OBJ files.

Fix export issues for some elements in Revit 2026. 

Fix incorrect transforms for some exported subcomponent elements. 

Support for Mars data added to imagery-tiler, terrain-tiler, and quantizedmeshto3dtilesconverter.

point-cloud-tiler handles PLY files with any number of comments now.

point-cloud-tiler fixed issue where in certain cases when using draco compression with .pnts output could result in batch table data not being written.

imagery-tiler added TileJSON format support.

Check the changelogs for the full list of updates for Cesium ion SaaS and Cesium ion Self-Hosted.

For years, Propeller has been using CesiumJS to visualize drone-captured surveys of earthworks, quarries, mines, and waste management sites. Propeller recently got an upgrade when its team incorporated Cesium’s Design Tiler into the platform. 

For Propeller’s users, streaming IFC files as 3D Tiles means a faster, higher quality experience. They can analyze large files, like viewing cross sections of this IFC model. Courtesy Propeller.

Large scale projects require massive coordinated efforts from many people across many businesses, whether that’s grading hundreds of acres for new suburbs, laying thousands of miles of pipelines, or hauling tens of thousands of cubic yards of dirt from quarries. Even minor errors can be costly, and sharing a common ground between the many teams involved is critical.  

Propeller supports users working at large scale sites: quarries, mines, earthworks, waste management, etc. Courtesy Propeller.

For users like these, Propeller provides the tools to perform frequent accurate drone surveys, process that data, and visualize the site. When this isn’t frequent enough, DirtMate brings near-real-time data into the mix. Propeller’s platform gives users an intuitive 3D model that’s recent, realistic, and accurate, so they can spot safety hazards, track earthworks progress, and measure reality against design plans to make calculated operational decisions around material, labor, and equipment needs. 

Propeller’s users perform drone scans of their sites and then visualize those scans in 3D with CesiumJS. Now, with support for IFC files as 3D Tiles, they’re able to compare designs against reality. Courtesy Propeller.

The heart of Propeller’s visualization is CesiumJS. Propeller was an early adopter of Cesium: for years they’ve been using Cesium’s open source software to visualize drone-captured surveys of earthworks, quarries, mines, and waste management sites. Propeller’s commitment to accuracy—reliably achieving geospatial accuracy within 3 cm—has made their platform trusted by professional surveyors. The sophisticated tooling they’ve brought into their platform, which they put into a simple form factor, makes advanced analysis accessible to anyone involved in the project, regardless of technical background.

Propeller users can see their project in accurate 3D. For pipelines, electric utilities, railways, and other long linear assets, the geospatial coordinate system in CesiumJS renders designs aligned with real-world curvature. Courtesy Propeller.

Propeller recently got an upgrade when its team incorporated Cesium’s Design Tiler into the platform. The Design Tiler takes complex BIM data, like the Industry Foundation Classes (IFC) open standard, and tiles it into 3D Tiles for efficient streaming. These improvements allow users to look at much more granular data, on a much larger scale, faster than ever.

"Integrating the Cesium Design Tiler into our platform has revolutionized our support for BIM for all our customers.” 

—Harry Morris, software engineer, Propeller

3D Tiles has made IFC files from one of the most cumbersome to one of the most performant file types on Propeller’s platform. This video shows first the original solution of an IFC model loaded in Propeller’s platform and then the same model tiled as 3D Tiles. Courtesy Propeller.

A visual of IFC files loaded in Propeller's platform before 3D Tiles and after tiling.

IFC files in Propeller before and after 3D Tiles

The tiling pipeline transformed IFC files—once cumbersome and limited to 300 MB—into one of the most performant file types on Propeller’s platform, which now supports IFC files up to 2 GB. This 6–7x increase in file size capacity came with dramatic performance gains. 

“For our customers with the most demanding projects, framerates are up anywhere from 20% to 170%, and memory usage has been drastically reduced, with our most demanding models seeing a 60x reduction in memory usage and similar reductions in data transfer," said Harry Morris, Propeller software engineer.

When streamed as 3D Tiles, IFC files retain all metadata: an inspector can click on an individual bolt in a bridge and see its size, manufacturer, the contractor responsible for installing it—whatever the users have uploaded. Courtesy Propeller.

Bringing Cesium’s Design Tiler into the platform took the Propeller team less than two months to roll out to full production deployment. When they released the new feature, the team also retroactively tiled all previously uploaded IFC files, delivering immediate value to their users without requiring re-uploads or workflow changes:

"Integrating the Cesium Design Tiler into our platform has revolutionized our support for BIM for all our customers. Cesium's expertise in the 3D Tiles open standard has delivered us a solution with greatly improved performance, accuracy, and feature richness—we've delivered highly requested features and given our construction customers confidence in an easy-to-use pipeline, revolutionizing their workflows," Morris said.

Explore how you can use 3D Tiles to make your datasets more accessible with Cesium’s 3D tiling pipelines.  

Propeller Streaming Complex IFC Files for Construction with 3D Tiles 

Last year, Cesium released the first readily accessible, comprehensive 3D Tileset of the Moon. Today, we are excited to introduce Cesium Mars. Cesium Mars is a 3D Tileset to enable developers, researchers, and mission planners to create visualizations and simulations of the Red Planet. 

Cesium Mars comprises precise terrain captured by NASA and the European Space Agency (ESA), paired with colorized imagery. Cesium Mars is provided as 3D Tiles, the Open Geospatial Consortium (OGC) community standard created by Cesium. 

Enabling a digital ecosystem for planetary data 

A wealth of Mars data has already been captured by orbital and rover missions. The seven man-made satellites from five space agencies currently orbiting Mars and the two active rovers on Mars, NASA's Curiosity and Perseverance, send home more data every day through the Mars Relay Network. In addition, two space agencies are actively planning manned missions to Mars, and several additional agencies are planning for continued robotic exploration and orbital analysis. Cesium Mars lowers the barrier to entry for studying existing and new Mars data and for leveraging this ever-increasing data as part of the planning, execution, and analysis for Mars programs. A precise, easily accessed model of Mars supports the space community in its efforts toward an open, interoperable platform for a digital ecosystem supporting data captured beyond Earth.  

Mars viewed from space. Data for Cesium Mars from the Mars Orbiter Laser Altimeter (MOLA) on the NASA Mars Global Surveyor (MGS) and the High-Resolution Stereo Camera (HRSC) on the ESA Mars Express. 

The Gale Crater, visualized with Cesium Mars.  

Cesium Mars draped with a Mars MOLA elevation map from NASA's Scientific Visualization Studio. 

The launch of Cesium Mars comes at the right time when there's a need for planetary-scale digital twins and simulation for the next leap in human and robotic spaceflight. Buendea intends to leverage Cesium Mars within NASA’s XR Operations Support System (XOSS) to expand our digital twin environments beyond the Moon and into Mars. By combining Cesium’s geospatial platform with our high-fidelity interactive environments in Unreal Engine, we are creating a foundation for next-generation mission planning, astronaut training, and real-time operations in support of these efforts.  

Cesium Moon in XOSS in VR mode. Courtesy Buendea and NASA’s Advanced Operations Concepts Lab.

Since Cesium Moon’s release, the dataset has become a foundation for digital twins for mission planning, for training simulators, and for games. We anticipate similar uses for Cesium Mars. 

Space architects designing for the Moon and Mars are already employing workflows like those used for infrastructure on Earth. Cesium Mars enables those processes to converge, as users leverage functionality available for our open source runtime engines. Use Cesium Mars as a base layer to place models of habitats, vehicles, or rovers on the Martian surface. If you have your own high-resolution terrain data, clip and inset it so your projects can live within a single continuous 3D surface rather than multiple separate project areas. Share realistic, interactive 3D design visualizations with large, dispersed teams to facilitate review, highlight design alternatives, and present designs as they change over time. 

NASA’s Perseverance rover on Mars, viewed with Cesium for Unreal in Unreal Engine.

Cesium Mars is provided as 3D Tiles using the IAU_2015:49901 CRS and is composed of the ⁠Mars Viking Colorized Global Mosaic 232m imagery and ⁠Mars MGS MOLA - MEX HRSC Blended DEM Global 200m for terrain, along with higher resolution data for the Jezero and Gale craters. Cesium’s visualization engines—CesiumJS, Cesium for Unreal, Cesium for Unity, and Cesium for Omniverse—are ready for you to start building with Cesium Mars using an accurate representation of the planetary ellipsoid for measurement and analysis. Cesium Mars is available via Cesium ion SaaS and Cesium ion Self-Hosted.

This CesiumJS Sandcastle provides an example of how to include Cesium Mars in a web app. A sample level in Cesium for Unreal provides an example for how to get started with development using Cesium Mars in Unreal Engine. 

As lunar and interplanetary missions proceed, the need for an open, interoperable, standards-based platform to support simulation, planning, and analysis has never been greater. Cesium Mars, along with Cesium Moon, is a step toward this vision, offering a precise and scalable foundation. 

To access Cesium Mars, sign up or sign in to Cesium ion, locate the asset in the Cesium ion Asset Depot, and add the tileset to your Asset Library. Then start building! Create Cesium Stories, incorporate the data into a CesiumJS web app, or build advanced visualizations with 3D creation engines such as Unreal Engine. 

In August, we attended SIGGRAPH, MODSIM World, TAK Offsite, and P3DL; celebrated YouthMappers’ grant-funded 3D course; and introduced iTwin Platform to Cesium developers.  

Follow the tutorials to kickstart your use of iTwin Platform with Cesium.

CesiumJS 1.133 is now available. Highlights of the release include:

Added Ellipsoid.MARS for use with Mars terrain and imagery. #12828 

Added spherical harmonics support for Gaussian splats, supported with the SPZ compression format. #12790 

Added support for the EXT_mesh_primitive_restart glTF extension. #12764 

Fixes issue where a Gaussian splat tileset would be rendered even if out of current camera view. #12840 

Removes the minimum tile threshold of four for WMTS. #4372 

Fixed a crash when loading PNTS (point cloud) data that contained a batch table without a binary part. #11166 

For more details, see the changelog. We also welcome your thoughts on the new version of Sandcastle.

The path of NASA's Mars Perseverance rover rendered in CesiumJS. Data courtesy NASA, European Space Agency, and the USGS Astrogeology Science Center.

Cesium for Unreal v2.19.1 is now available. Highlights of this release include:

Added assets for the IAU2015 Moon and Mars ellipsoids under Content. 

Added a missing CesiumRuntime.h include in CesiumPropertyAttribute.cpp that broke compilation in v2.18.0 on Windows. 

Fixed a bug in FCesiumVectorStyle that made polygons use the color mode from LineStyle instead of PolygonStyle. 

Fixed a bug that could cause a crash when undoing and redoing deletion of a CesiumRasterOverlay component. 

Fixed a crash in UCesiumLoadGeoJsonDocumentFromIonAsyncAction that occurred when a Cesium ion server was not specified. The action will now fall back to the default ion server, similar to ACesium3DTileset. 

Fixed a bug that could cause incorrect ellipsoid parameters to be used when loading a level. 

In addition to the above, this release updates cesium-native from v0.50.0 to v0.51.0. See the changelog for a complete list of changes in cesium-native. 

Cesium for Unity v1.18.0 is now available. Highlights of this release include:

Added option to ignore the KHR_material_unlit extension to force default lighting on tilesets. 

Fixed a bug where CesiumCreditSystem did not accurately track the loading progress of images, which could result in missing credits. 

Fixed a bug where CesiumPointCloudRenderer could dereference a null CesiumPointCloudShading on Cesium3DTileset. 

Check the Cesium for Unity changelog for the full list of updates, and follow our Cesium for Unity tutorials to get started. 

Coming soon: Cesium Mars. Stay tuned for the full announcement later this week. 

Fixed an issue with texture size estimation for tiles, which could negatively impact performance and cause tileset trees to be deeper than necessary. 

Fixed an issue that would cause some tile textures to have an invalid sampler index. 

Fixed an issue that could cause a crash or invalid material references when loading a glTF file with multiple primitives that reference the same material. 

Fixed an issue where materials would be incorrectly handled when loading multiple glTF files. 

Fixed an issue where translucent IFC materials were being considered opaque. 

Fixed an issue where an excessive number of tiles would be generated when processing nodes with a single point. 

"The most brilliant engineering designs, no matter how well crafted, remain unrealized until their vision can be clearly communicated and understood by all stakeholders," said Wayne Sullivan, visualization lead at infrastructure engineering and consulting firm Parametrix. This philosophy drives Parametrix’s Advanced Visualization team: a group of software engineers and 3D artists using Cesium for Unreal, Cesium ion, and Google Photorealistic 3D Tiles to enhance public outreach and help win project bids by making complex designs easier to understand. 

High-fidelity, photorealistic renderings of a proposed design helped the City of Las Vegas secure a $25 million grant for the Stewart Avenue Complete Streets project. Courtesy Parametrix.

We’re excited to share that we’ve started to bring iTwin Platform APIs to Cesium, aiming to provide a complete platform for digital twin experiences for the built and natural environment, shaped by the community's real-world insights and continuing feedback. iTwin APIs power digital twin solutions used in architecture, engineering, and construction (AEC) apps—and across industries—to design, build, and operate infrastructure around the world. By exposing these APIs through Cesium, we want to enable Cesium and iTwin developers to build powerful applications with full geospatial context and advanced infrastructure digital twin workflows. 

To start, we’ve created a tutorial series for Cesium developers trying iTwin Platform. This initial set of tutorials helps you build a CesiumJS app with reality data from Cesium ion and design models from iTwin. We want to hear about your experience using iTwin APIs and what APIs or capabilities you would like to see coming to Cesium first. 

The iTwin tutorial series helps you kickstart your use of iTwin with Cesium.

iTwin Tutorial Series

iTwins are digital representations of physical infrastructure assets—such as buildings, bridges, or other structures—that can be visualized, analyzed, and interacted with throughout the life cycle of a project. Bentley Systems' open-core iTwin Platform provides the tools and APIs to create, manage, and collaborate on these digital twins within a geospatial context, enabling clearer insights and better decision-making. 

Using the iTwin Platform with CesiumJS to search for and visualize all solar panels in the design.

Visualizing solar panels with iTwin Platform and CesiumJS

What iTwin Platform introduces to the Cesium community 

The integration of iTwin with Cesium unlocks new possibilities for developers, engineers, and decision-makers: 

Broad design format support: Work with a wide range of formats from popular AEC applications, including MicroStation, AutoCAD, Navisworks, Revit, Rhino, and more. The full list of supported formats is here. 

Design models as 3D Tiles: iTwin automatically converts design data into 3D Tiles with efficient level-of-detail management. 

Integrated design history: Track the progression of your designs over time and compare alternative design options. 

Rich metadata access: Query your design model’s metadata using ECSQL, enabling powerful insights and filtering capabilities.  

Tools applicable across industries: Simple APIs for powerful workflows, such as clash detection, change detection, forms, and issue tracking.

Visualizing clashes detected by the iTwin Clash Detection API.

Getting started with iTwin tutorials for Cesium 

To help you hit the ground running, we’ve released the first series of tutorials that walk you through the basics of using iTwin with Cesium: 

Adding Design Context Using Cesium ion: Learn how to create a geospatially accurate context model for your project. 

Visualizing iTwin Design Data in CesiumJS: Visualize your iTwin in a CesiumJS application and explore your design in 3D. 

Visualizing iTwin Design Options in CesiumJS: Switch between multiple design options with a simple user interface. 

Toggling design options in the tutorial app.

iTwin + Cesium: Toggle Design Options

What’s next—and how you can help shape it 

We’re just getting started in introducing what’s possible with iTwin and Cesium. In future tutorials, you can expect to see examples of querying metadata from your iTwin in Cesium applications, interacting with elements through spatial searches or SQL-like queries, and navigating directly to specific components in your model. We'll also explore how to combine multiple iTwins within a single scene and how to incorporate advanced iTwin APIs like clash detection, change tracking, and issue management, bringing even more engineering intelligence into your Cesium-based geospatial applications.

We want to hear from you about which use cases we should prioritize as tutorials.

Are there specific workflows, formats, or practical examples you’d like us to cover? As you build tools for design review, infrastructure monitoring, or digital twin dashboards, your input will help shape where we go next. 

Share all of your requests with us on the community forum.

Introducing iTwin Platform to Cesium Developers

本記事は、2025年6月23日にShehzan Mohammedによってリリースされた記事を、今回日本語化したものです。

CesiumがBentley Systemsの一員となってから約9ヶ月が経ちました。この間、私たちはBentleyの技術を深く理解し、それをコミュニティにどのように還元できるかを探ってきました。今回、初のCesium Developer Conferenceを機に、皆さまからのご意見を広くいただくべく、iTwinによるデジタルツイン機能をCesiumに統合するためのロードマップ案を共有します。

当社のミッションは、「人工環境と自然環境が融合する3D地理空間体験をコミュニティが自由に構築できるようにすること」です。3D Tilesの活用は着実に拡大しており、Google、Esri、Ansys、FOX Sports、 The Weather Company、 GE Aerospace、 MathWorks、 CAE、 Trimble、 そして DroneDeployなど、多くの企業に採用されています。Cesium ionも月間30,000人以上のアクティブ・ユーザーから3億回以上のリクエストを受けています。Cesiumはオープンソースへの継続的な投資を通じて、コミュニティの拡大とともに、CesiumJS、Cesium for Unreal、Cesium for Unity、Cesium for Omniverseといったランタイムエンジンを活用し、新たな3D地理空間アプリケーションの迅速な開発を可能にし続けています。

また、現在以下のようなニーズに限りないほどの強い需要がある状況です。

多様な入力形式に対応する高性能なタイル変換パイプライン

より効率的なエンドツーエンドのワークフロー支援ツール

Bentleyの一員として、私たちはBentleyの最先端技術を活用し、3D地理空間コミュニティ全体を根本的に強化・拡大していく機会を得ています。

Cesiumはこれまでも常にコミュニティの声を指針としてきました。今回のロードマップに掲げたテーマについても、皆さんのフィードバックをぜひお寄せください：

リアリティモデリングおよびAI解析サービスのCesium ionへの統合

iTwinプラットフォームAPI／サービスをCesium経由で利用可能にする

データを3D Tilesとして最適化するための柔軟に構成可能な(composable)タイル処理パイプラインの導入

新3D Tilesへの拡張やプラットフォーム統合に対応したCesiumランタイムエンジンの進化

高品質な整備済みの地形・都市・惑星データ提供をさらに拡充

再利用可能でカスタマイズ可能なUXツールとウィジェットの提供

学習教材の充実と継続的なメンテナンス強化

3D Tilesは、Cesiumプラットフォームの中核を成す存在です。
 プラットフォーム全体とコミュニティを貫く「軸」として、エンドツーエンドの一貫したワークフローを形づくり、開発者が卓越したアプリケーションを構築できる力を与えてきました。私たちは現在、コミュニティから長年にわたって求められてきた複数の新機能を3D Tilesに追加しようとしています。

あらゆる業界のユーザーが、これまでになく高解像度かつ高頻度で、さまざまな種類のデバイスを使って環境をキャプチャするようになっています。
 このような指数関数的なスケーリングの進展は、空間的整合性を保ちつつ、時間とともに更新される3D地理空間データを、より効率的に保存・ストリーミング・可視化・分析できるよう、3D Tiles仕様を拡張する絶好の機会を生み出しています。

時系列対応 3D Tilesは、3D Tilesに時間軸の概念を効率的に組み込むための拡張仕様です。
 これを使えば、1つのタイルセットを時間の経過とともに更新でき、各キャプチャで変更があった部分のタイルだけを差分更新することが可能になります。

もはや、更新のたびに完全な新しいタイルセットを作り直す必要はありません。
 1つのタイルセットに変更履歴を保持しながらインクリメンタルに更新できるため、再生成の効率も速度も大幅に向上します。

この仕組みは、時間に沿った変更点の分析や、タイルセットと設計モデルの時系列比較といった分析ワークフローにも活用可能です。 詳細や意見交換はGitHubのIssueページ にて。

時系列対応3Dタイルを使えば、1つのタイルセットを時間の経過に沿って更新でき、各キャプチャごとに影響を受けたタイルだけを差し替えることが可能になります。

さらに現在、BentleyのSYNCHRO 4D Schedule APIと連携し、時間変化に対応する3D Tilesetを生成する手法も検討しています。
 SYNCHRO 4D Schedule APIは、時間軸を含んだプロジェクトデータをWeb経由で取得・活用できる仕組みを提供しており、この連携により、時間的に変化するジオメトリと業務インテリジェンスとの融合が実現されます。

ボクセルによるボリューム（体積）データ表現

地質学、海洋学、気候科学など、連続的なデータ表現が求められる多くの分野において、ボリュームデータは現実世界を可視化する重要な手法となっています。

3D Tiles仕様は、大規模かつ多様なデータを最適に表現できるよう、さまざまなジオメトリ形式を柔軟にサポートするよう設計されています。

3D Tilesにボクセル（体積要素）を3Dペイロードとして導入することで、地球規模のボリュームデータを扱うアプリケーションの構築が可能になります。

現在私たちは、glTFをボクセルのペイロード形式として活用しながら、楕円体・直方体・円柱型のボクセルを3D Tiles仕様の出発点として開発を進めています。

これにより、地中・地表・大気・宇宙空間など、さまざまな領域における体積データの表現と活用が可能になります。

 詳細はVoxels拡張仕様のドラフトをご覧ください。

CesiumJSでボクセルを使用して、メキシコ湾の海水温を表現

3D Gaussian Splats（3Dガウシアン・スプラット）

近年のコンピュータグラフィックス分野で最も注目を集めている技術の一つが、3D Gaussian splats（3Dガウシアン・スプラット）です。

高解像度のリアリティメッシュや点群は、自然環境のデジタル化に非常に優れた手段ですが、ケーブル、樹木、鉄棒、通信塔といった一部の情報の捉え方には限界があります。

その点、3D Gaussian splats は、極めて高い再現性を保ちつつ、視覚的なディテールを精緻に再現できる優れたソリューションであることが証明されています。

私たちは現在、3D Tilesのペイロードタイプのひとつとして3D Gaussian splatsを追加することに取り組んでいます。

これに伴い、iTwin Captureで使われている技術をCesium ionに統合し、Gaussian splatsのモデリングとタイル化を可能にする計画も進めています。

さらに、glTFコミュニティと連携し、glTFに対応した3D Gaussian splat拡張仕様の策定にも取り組んでいます。

3D Tilesで表現した3D Gaussian splats（左）と、Reality Mesh（右）との比較。特にワイヤー、パラボラアンテナ、支柱、足場といった細長い形状の表現力に注目。

ジオメトリの強調表示とカスタマイズ機能

Bentley Systemsは、3D Tilesの初期採用企業の一つであり、インフラの詳細モデルに必要な3Dビジュアライゼーション機能を強化するための拡張機能を独自に開発してきました。CesiumがBentleyの一員となったことで、こうした拡張機能やランタイム実装を統合し、コミュニティ全体に提供する機会が生まれています。現在私たちは、以下のような表現機能の追加に取り組んでいます：エッジ表示とワイヤーフレーム、線の描画とマテリアル設定（線幅、パターン、色など）、ポイントサイズのカスタマイズ、注釈表示用のビルボード、その他多数の拡張機能

glTFの EXT_mesh_primitive_edge_visibility 拡張を用いてエッジをハイライト表示した、iTwin.js 上の3D Tilesetの例。

暗黙面を使うことで、パイプ、柱、梁、ボックス、球体、円柱といったパラメトリック（数式で定義される）形状を表現できるようになります。

さらに、他の暗黙面と組み合わせて使用することで、三角形メッシュでは表現が難しい複雑なジオメトリの作成も可能になります。

この手法の主な利点は、ストレージ効率の向上と解析精度の向上にあります。

というのも、これらの面は三角形メッシュによって近似されるのではなく、数学的に厳密な形状として保持・描画されるため、ジオメトリ的な誤差が発生しないのです。

私たちは現在、Microsoftが導入した glTF の暗黙的形状拡張（implicit shapes extension）を基盤として、この機能の実装を進めています。

3D TilesにおけるBISセマンティクス（意味情報）の対応

3D Tiles仕様は、ジオメトリに属性や分類情報を付加する詳細なメタデータをサポートしています。このメタデータ仕様は、さまざまな粒度・産業・ユースケースに対応したセマンティクス（意味情報）を付与できるように設計されています。

とくに建築・エンジニアリング・建設（AEC）業界においては、オブジェクトのクラス（種類）に適用可能な標準化された分類語彙（タクソノミー）の整備が求められています。こうした分類の意味を標準語彙として明示的に定義しておくことで、AI によるデータ検証、データ連携、分析、インタラクションなどにおいて、後続処理に有効活用できるようになります。

たとえば、ユーザー独自スキーマ内の「あるクラス」が 「ドア(door)」 を表している場合、それを IFC（Industry Foundation Classes）、Uniclass、あるいは BIS などの標準語彙に対応づけておけば、他のデータ利用者が容易に扱うことが可能になります。

BIS（Base Infrastructure Schema）は、インフラ工学分野のフェデレーションデータをモデリングするための分類体系、データ構造、関係性を表すモジュール式スキーマ群です。これらの BIS スキーマはレイヤー構造で編成されており、大部分は BIS エコシステムにおける標準的なルールやパターンを定義しています。標準的なBISスキーマに組み込まれたセマンティクス（意味情報）は、iModelの枠を超えたさまざまな分野でも有用であることが証明されています。

現在私たちは、クラス（分類）を扱えるようにするためのメタデータセマンティクス拡張を3D Tiles向けに開発しており、BISをはじめとする同様のセマンティック語彙体系を適切にエンコードできるように取り組んでいます。

Cesium ion へのリアリティモデリングと AI 解析サービスの統合

Cesium が Bentley の一員となったことによる最もエキサイティングな成果のひとつが、リアリティモデリングと AI を活用した解析サービスを Cesium ion に統合できるようになったことです。

「写真をアップロードして 3D モデルを生成したい」という声はこれまでにも多く寄せられており、私たちは、業界をリードする iTwin Capture のリアリティモデリング技術をCesium ion に統合することで、ついにそれを実現できることをとても嬉しく思っています。

写真から 3D Tiles を生成し、さらにそれをレンダリングや解析へとつなげる統合されたエンドツーエンドのパイプラインを通じて、リアリティデータを用いた 3D 地理空間アプリケーションにおいて、最高水準のユーザー体験を提供することを目指しています。

さらに、3D Gaussian splats の技術進化により、写真からスプラット形式の 3D タイルセット を生成するワークフローもサポートできるようになります。

iTwin Captureで使われているリアリティモデリング技術を活用し、Cesium ionにReality Modeling機能が加わったことで、写真データから直接3D Tilesを作成できるようになりました。

Terradepth Processes Seafloor Data at the Edge with Cesium ion Self-Hosted

Cesium Releases in October 2025

Propeller Streaming Complex IFC Files for Construction with 3D Tiles

Lift Off to Cesium Mars

Cesium Releases in September 2025

Introducing iTwin Platform to Cesium Developers

人工環境と自然環境の統合に向けたCesiumの技術ロードマップ