3DISE conference

Michal Gula & Tomas Barnas
  • 0.0 (0)
  • Entrepreneurship
  • Weekly

After a successful debut in 2025, we’re excited to welcome you to the 2nd edition of 3DISE Conference in Prague! 3DISE is the premier gathering for industry leaders, experts, and enthusiasts shaping the future of photogrammetry, laser scanning, surveying, Gaussian splatting, immersive technologies, 3D printing, and beyond.

  1. 19h ago

    Gaussian splatting is not enough: 95% still use point clouds | Blurry

    Martin Piala is co-founder of Blurry, a London-based startup building a workspace for reality capture data: point clouds, Gaussian splats, images and videos in one place, in the browser. Blurry started a few months before 3DISE, and INTERGEO 2026 was its first large trade show, straight in as an exhibitor. We talk about why Blurry stopped being only a Gaussian splat viewer after seeing that 95% of its users still work with point clouds, why it partners with Teleport by Varjo for processing instead of building everything itself, and why splats in VR headsets are a can of worms it hasn't opened yet. Then AI: why Blurry hasn't shipped a half-baked AI feature, why Martin sees the data format, not the model size, as the limiting factor, and whether there is enough 3D training data. He also shares the funding picture: pre-seed from two VC funds, around three years of runway, and a seed round in the millions planned within six to nine months. Recorded live at INTERGEO 2026 in the 3DISE podcast studio. Hosts: Michal Gula and Tomáš Barnáš. Guest: Martin Piala, Blurry. Chapters:0:00 Blurry: a Google Drive for reality capture data0:35 Why Gaussian splatting alone is not enough1:15 From NeRFs to Gaussian splatting2:05 Viewing, storing and now processing2:36 95% of users still work with point clouds3:59 Can one company do processing and viewing?4:58 Own processing for 360 cameras5:40 No hardware: partnering with manufacturers6:01 Gaussian splats in VR headsets7:13 Why there's no half-baked AI yet8:14 Can AI spot BIM vs scan discrepancies?9:39 Frontier AI models in a small startup10:39 AI doesn't understand point clouds11:44 Is there enough 3D training data?12:35 Is a point cloud just text?14:33 Pivot or course correction15:12 Pre-seed funding and the next round16:11 London HQ and a seed round in the millions17:11 Wrap-up

  2. 2d ago

    SLAM will never replace terrestrial scanning – Aden Zhang, XGRIDS

    Aden Zhang is Sales Director for Asia Pacific at XGRIDS, with a survey background and four years at the company. At INTERGEO 2026 he brought the L3, the follow-up to the L2, which per Aden shipped around 4,000 units last year. We go through what changed: a built-in screen instead of the phone mount, a double base plate for closing SLAM loops, a larger camera sensor for Gaussian splatting, and the accuracy figures XGRIDS markets from its white paper. Then the uncomfortable part. Aden says SLAM will never replace terrestrial scanning, that newcomers who scan 80 minutes without loop closures get poor accuracy, that Gaussian splatting is not for cadastral survey, and that making LCC more closed last year was a mistake XGRIDS has since corrected. Open question: should a SLAM scanner ever be someone's first scanner? Recorded live at INTERGEO 2026 in the 3DISE podcast studio. Hosts: Michal Gula and Tomáš Barnáš. Guest: Aden Zhang, XGRIDS. Chapters:0:00 Aden Zhang, XGRIDS sales in Asia Pacific1:16 80% of new products are handheld LiDAR2:20 Why the double base plate matters for SLAM4:27 A screen instead of the phone mount5:22 Bigger camera sensor for Gaussian splatting6:54 SLAM accuracy and noise claims8:23 Drone mounting, but no aerial LiDAR9:44 Splats are not for cadastral survey10:54 Should newcomers buy SLAM?12:33 80 minutes without loop closures14:31 What comes after PortalCam15:18 Why scanners skip better cameras18:30 16 vs 32 channels: density, not accuracy19:37 Closing LCC was a mistake22:15 Chinese vendors and the rest of the world23:38 Wrap-up

  3. 3d ago

    Cesium open source: why giving it away became the sales funnel

    Danny Williams was part of Cesium, the small company behind the open source 3D geospatial platform, when Bentley Systems acquired it. Before Cesium he worked at Epic Games, when Epic acquired the RealityCapture company, so he has seen acquisitions from both sides of the table. We talk about Patrick Cozzi's decision to make Cesium open source when everybody asked how it would make money, why that open source code now feeds the commercial business, and why Cesium chose Bentley. Then Bentley itself: 40 years in software, an early idea of a 3D modelling tool for furniture, a catalog of software tools in the hundreds, a consumption-based E365 license and a new mobile app for reality capture. Plus the ContextCapture, iTwin Capture and RealityScan naming confusion, and aerial photogrammetry from balloons before drones. Open question: when a company buys instead of builds, who makes all those tools work as one? Recorded live at INTERGEO 2026 in the 3DISE podcast studio. Hosts: Michal Gula and Tomáš Barnáš. Guest: Danny Williams, Cesium (Bentley Systems). Chapters:0:00 Danny Williams, from Cesium to Bentley0:43 What changes after an acquisition2:15 Why Cesium went open source3:03 Open source as a funnel, not a risk3:40 Why Bentley was the right home5:08 Where is my Bentley car?5:59 From furniture tool to infrastructure7:08 Platform: where new tech is incubated7:47 Hundreds of software tools8:47 Buy, don't build: Bentley's model10:38 One environment for disparate data11:19 E365: one license, pay per use12:32 A mobile app for reality capture14:10 From ContextCapture to iTwin Capture15:14 RealityScan naming and the Epic years16:33 Balloons and film cameras before drones18:02 Wrap-up

  4. 5d ago

    Trimble Connect: point clouds, 360 images and AI search in one place

    Gregor Willhauck has been at Trimble since 2010, when Trimble acquired the eCognition business of Definiens, the image analysis startup founded by Nobel laureate Gerd Binnig. Since then he has worked in product management on photogrammetry and laser scanning software and on rugged handhelds, and he now runs go-to-market for cloud for field systems, which at Trimble means Trimble Connect. We talk about how a BIM collaboration platform became the place for point clouds and 360 imagery, how a prefab wall maker in Canada checks rebar against a scan before the wall arrives, Gaussian splats for UAV data (still in development), AI search over project data, and the ST30 with a laser in the pole that measures the prism height. Then the uncomfortable ones: why Trimble has no SLAM scanner, why Scandinavia uses reality capture in the cloud more than the US, and Gregor's answer to which TLS to buy: the hardware is, in a way, exchangeable. Recorded live at INTERGEO 2026 in the 3DISE podcast studio. Hosts: Michal Gula and Tomáš Barnáš. Guest: Gregor Willhauck, Trimble. Chapters:0:00 First INTERGEO in 2000 with eCognition0:42 Joining Trimble through an acquisition1:22 Definiens: a Nobel laureate and two exits4:21 From software to rugged hardware5:45 Trimble Ventures and startups7:41 Skunk works: a modular hardware platform8:57 Cloud for field systems: Trimble Connect10:06 Why architects use Trimble Connect11:12 Scan vs. model before the wall arrives12:52 Gaussian splats for UAV data13:35 Splats are still in development14:14 AI search instead of manual filters16:11 What usage data shows about customers17:15 No MCP yet: AI agents on project data17:59 Connecting the physical and digital19:08 ST30: a laser in the prism pole20:46 Harder than rocket science21:59 Why Trimble has no SLAM scanner24:15 Scandinavia ahead of the US in the cloud27:29 Which TLS to buy? The ecosystem28:40 How SketchUp came from Google29:28 Why architects still use SketchUp31:12 Wrap-up

  5. Oct 23

    SLAM scanner range: 70 m on paper, 50 m of good data | FJD

    Eduard Chiru has spent two and a half years at FJD, after getting to know its SLAM scanners at a surveying company in Bucharest that resold them. He spends about 30% of his time travelling to shows, 3DISE included. FJD brought five scanners to INTERGEO, from entry-level devices at around 3 cm accuracy to the T1, which works as a handheld SLAM scanner and, on a tripod, as a TLS with 4 mm at 10 m. We talk about what the X1 does for site managers (tasks from the cloud, AI floor plans, voice notes translated by AI), why TLS mode is an algorithm switch and not a hardware one, and the honest numbers: on the X1, a 70 m range gives you about 50 m of good data day to day, per Eduard. Then the speed limits for walking, car and drone mounts, what happens when you scan from a scooter in handheld mode, and how FJD wants to stand out when most SLAM scanners look the same. Recorded live at INTERGEO 2026 in the 3DISE podcast studio. Hosts: Michal Gula and Tomáš Barnáš. Guest: Eduard Chiru, FJD. Chapters:0:00 A contact zoo at INTERGEO0:48 From a Bucharest reseller to FJD1:39 Six or seven shows a year2:40 Five new scanners at INTERGEO3:44 The X1: tasks sent from the cloud4:52 Automatic 2D floor plans with AI6:15 Voice notes translated by AI7:38 T1: SLAM and TLS in one device8:37 TLS mode is an algorithm switch10:13 One or two minutes per station11:05 How much of the range is usable?12:54 Drone, car and robo-dog mounts13:13 Speed limits from 1 m/s to 20 km/h14:15 Scanning from a scooter15:32 More SLAM companies every year17:36 AI beyond the chat window18:16 Wrap-up

  6. Oct 22

    Point clouds, Gaussian splats and BIM in one scene | NUBIGON

    Bert Azizoglu is the CEO and co-founder of NUBIGON, desktop software for rendering reality capture data. Its users are mostly service providers and scanning professionals who need to explain their data to clients. INTERGEO 2026 was NUBIGON's first trade show with its own booth. We talk about why NUBIGON stayed offline while others build web viewers (Bert says an online application is coming, and some clients in critical infrastructure prefer data that never leaves their side), about hybrid scenes that put point clouds, Gaussian splats, meshes, CAD and BIM models in one place so they validate each other, and about context: why renders usually show a design in a vacuum when most of Europe is already built. Bert's sharpest point: even when employees look at the data, decision-makers rarely do. And a reality capture dataset shows one moment in time, so how do you analyze sun exposure on buildings that already exist? Recorded live at INTERGEO 2026 in the 3DISE podcast studio. Hosts: Michal Gula and Tomáš Barnáš. Guest: Bert Azizoglu, NUBIGON. Chapters:0:00 First INTERGEO with a booth1:22 Running a booth with a small team2:30 Who NUBIGON works for4:17 Sun exposure on existing buildings5:23 Offline software vs. web viewers7:08 Renders for decisions, not only marketing7:44 Point clouds, splats and BIM in one scene8:47 Data stays on the client side9:42 Designs are not made in a vacuum11:30 Clients never open the delivered data12:35 Decision-makers rarely look at the data13:50 Next: online viewer and night renders14:23 One moment in time vs. changing buildings15:40 Sun exposure mapping per surface15:59 Wrap-up

  7. Jan 22

    Accuracy Is Not Optional: What Most People Get Wrong About Reality Capture

    In this episode, we talk with Paul Burrows from Hexagon, a leading expert in reality capture and measurement technology, about what really matters in the 3D industry today.We dive deep into:Why accuracy and measurement are more important than everThe biggest mistakes non-surveyors make when using reality capture technologyHow personal branding impacts trust, credibility, and long-term career growthWhy technology alone will not save you if you don’t understand the fundamentalsHow professionals can stand out in a crowded 3D and reality capture marketThis conversation is highly relevant for professionals working with:laser scanning, photogrammetry, reality capture, surveying, digital twins, construction, AEC, and geospatial technologies.Whether you are a surveyor, engineer, reality capture specialist, or business owner in the 3D industry, this episode will help you think differently about accuracy, responsibility, and positioning yourself as a trusted expert.

  8. Jan 22

    Accuracy is a can of worms with SLAM with Stefan Hrabar from EMESENT

    In this episode with Stefan Hrabar, Co-Founder of  @Emesent  we go deep into SLAM accuracy, LiDAR trajectory estimation, RTK GPS integration, and Gaussian Splatting — without marketing hype.You’ll learn:Why accuracy is the wrong first question when evaluating SLAM systemsHow trajectory estimation defines point cloud qualityWhy RTK GPS works like thousands of ground control points (GCPs)The real role of Gaussian Splatting in modern reality capture workflowsHow fast mobile LiDAR scanning works in real city environmentsThis conversation is highly relevant for:Reality capture professionalsSurveyors & geodesistsAEC industry specialistsLiDAR & SLAM system users3D mapping, digital twins & mobile mapping teams📍 Discussed use cases include urban scanning, city mapping, indoor & outdoor SLAM, and mobile LiDAR workflows used across Europe and globally.🔍 Topics coveredSLAM LiDAR, Mobile Mapping, RTK GPS, Point Clouds, Gaussian Splatting, Trajectory Estimation, Reality Capture, 3D Mapping, Digital Twins, AEC Technology

See All (17)

About

After a successful debut in 2025, we’re excited to welcome you to the 2nd edition of 3DISE Conference in Prague! 3DISE is the premier gathering for industry leaders, experts, and enthusiasts shaping the future of photogrammetry, laser scanning, surveying, Gaussian splatting, immersive technologies, 3D printing, and beyond.

Information

  • Creator
    Michal Gula & Tomas Barnas
  • Years Active
    2025 - 2026
  • Episodes
    17
  • Rating
    Clean
  • Copyright
    © Michal Gula & Tomas Barnas
  • Show Website
    3DISE conference