IceCypress Technology’s Post

What powers #IceCypress spatial intelligence solutions? Our product ecosystem is designed to connect every stage of spatial data workflows — from modeling to analysis and deployment. • Modeling Software Mirauge3D · Rusa Delivering high-precision 3D reconstruction and rapid scene modeling • Application Systems DroneSwarm · TrackSight Enabling real-time onboard aerial analysis and high-speed target tracking • Cloud Platform Low-altitude 3D Platform Providing integrated data management and scalable deployment capabilities Together, these components form a unified system that transforms spatial data into actionable insights for real-world applications — helping make low-altitude space more efficient, secure, and intelligent. Feel free to connect with us to explore how our solutions fit your business needs. #SpatialIntelligence #3DReconstruction #DigitalTwin #AI #DroneTechnology #SmartCity #Geospatial #Infrastructure

At NXT BLD, we’re turning the messiness of the real world into structured intelligence that BIM, analysis and AI can use. On Day 1 Dr. Florent POUX will show how AI transforms raw LiDAR and point clouds into usable insight for digital twins and infrastructure workflows. He'll be followed by a dedicated geospatial track that will focus on placing buildings within their broader physical, infrastructural and environmental context. Rather than treating site and city data as a background layer, the session will emphasise geospatial as a core reference frame for design, analysis and automation with presentations from Esri, cityweft, and Infraspace On Day 2 we'll delve deeper with an expert reality modeling panel featuring Julien Moutte (Bentley Systems), Mike Deacon (Autodesk), Irene Radcliffe (Faro) and Dr Florent Poux - chaired by Robert Klaschka. Reality modelling stops being documentation. It starts becoming intelligence. #NXTBLD #RealityModelling #DigitalTwins #Geospatial #AEC #AI #pointclouds

New Post: Three‑Modal Fusion for Real‑Time Indoor Mapping and Semantic Understanding on Spatial‑Computing Platforms - — ## Abstract Indoor mapping and navigation are essential for augmented‑reality, robotics, and asset‑management applications within spatial‑computing ecosystems. Single‑modal approaches that rely exclusively on LiDAR or RGB cameras typically trade off between geometric accuracy, semantic richness, latency, and robustness in cluttered indoor spaces. This exploratory study proposes a tightly‑coupled three‑modal fusion framework that simultaneously processes \[…\] \[Source & Legal Disclaimer\] This is an AI-generated simulation research dataset provided by Freederia.com, released under the Apache 2.0 License. Users may freely modify and commercially use this data \(including patenting novel improvements\); however, obtaining exclusive patent rights on the original raw data itself is prohibited. As this is AI-simulated data, users are strictly responsible for independently verifying existing copyrights and patents before use. The provider assumes no legal liability. For future Enterprise API access and bulk dataset purchase inquiries, please contact Freederia.com.

How do you align a noisy, incomplete 3D scan with a perfect digital model? This is a fundamental problem across many 3D vision applications, including construction monitoring and robotics — but real-world point clouds are messy:   Noise, occlusions, missing regions, and inconsistent density make reliable alignment difficult. To support research in this space, we introduce PC2Model — an ISPRS benchmark for point cloud-to-model registration.   This work will also be presented at the XXV ISPRS Congress 2026(ISPRS 2026 Toronto). Instead of relying only on clean or limited datasets, PC2Model is designed to reflect reality: • Simulated LiDAR scans with realistic artefacts (noise, occlusions, mixed pixels, and density variation)   • Real-world scans for added complexity   • Ground truth transformations   • Standardised metrics: LOA (accuracy), LOC (coverage), and transformation error   The dataset supports both classical and learning-based methods and is designed as a resource the community can build on. 📄 Paper: https://lnkd.in/dWa2_ae2   📂 Dataset download: https://lnkd.in/dGjGREvf   🔧 Dataset repo: https://lnkd.in/dRRVXqEQ   ⚙️ Tools & simulation pipeline: https://lnkd.in/d-8WcPtq   We hope this benchmark helps move research towards more robust and realistic 3D registration methods.   We invite the community to use, evaluate, and build upon this resource. Collaborators:   Mehdi Maboudi, Said Harb, Jackson Ferrao, Kourosh Khoshelham, Yelda Turkan, Karam Mawas   Affiliations & support:   Technische Universität Braunschweig   ISPRS - International Society for Photogrammetry and Remote Sensing   AMC - TRR 277 #PointCloudRegistration #3DPointCloud #LiDAR #Photogrammetry #ISPRS #ComputerVision #DeepLearning #OpenDataset

Point Cloud to 3D Model Registration is a hot topic. Recently we published our new paper and dataset PC2Model in ISPRS, contributing to researchers developing better and faster deep learning models for point cloud to model registration.

Interesting study using the #HELIOS++ LiDAR simulator (https://lnkd.in/gkPSdSEu) for point cloud-to-model registration. ⬇️ Nice, Blender is definitely the most elegant option to create 3D scenes for LiDAR simulation. Stay tuned, with the upcoming HELIOS++version 3.0 and its completely new Python API it will become super easy to create your own plugins (Blender, GIS, CloudCompare, etc.) or to simply use the brand-new standard HELIOS++ Blender plugin developed by us. Looking forward to the talk in Toronto. 3DGeo Research Group Heidelberg | Hannah Weiser | William Albert

Inertial Labs has enhanced its geospatial capabilities with the latest PCMasterPro 1.16 update—bringing improved performance, accuracy, and efficiency to 3D terrain modeling workflows. “Having trusted data and end-to-end 3D processing and refinement across hardware and software is essential to ensuring accuracy, consistency and reliability from capture to visualization,” said Jamie Marraccini, Vice President, Inertial Labs, a VIAVI Solutions Company Products, VIAVI Solutions. “PCMasterPro, which supports the RESEPI product line, offers these capabilities – from tightly coupled inertial-based algorithms and reporting to locally referenced simultaneous localization and mapping (SLAM) generated point clouds. These tools enable professionals to make confident decisions, scale complex workflows and create digital twins that realistically reflect the real world.” #3DMapping #LiDAR #Geospatial #TerrainModeling #Innovation #Surveying #DigitalTransformation #powerelectronics #powermanagement #powersemiconductor https://lnkd.in/gaPj-fMQ

New Post: Temporal Consistency and Contrastive Multi‑Modal Fusion for 3D Scene Graph Generation in Urban Driving Environments - — ### Abstract We introduce a fully end‑to‑end framework for generating temporally consistent 3D scene graphs from synchronized LiDAR, RGB camera, and radar streams in urban driving scenarios. The proposed architecture couples a Transformer‑based temporal encoder with a multi‑modal contrastive learning module that aligns point‑cloud, image, and radar embeddings before graph decoding. We employ a \[…\] \[Source & Legal Disclaimer\] This is an AI-generated simulation research dataset provided by Freederia.com, released under the Apache 2.0 License. Users may freely modify and commercially use this data \(including patenting novel improvements\); however, obtaining exclusive patent rights on the original raw data itself is prohibited. As this is AI-simulated data, users are strictly responsible for independently verifying existing copyrights and patents before use. The provider assumes no legal liability. For future Enterprise API access and bulk dataset purchase inquiries, please contact Freederia.com.

Eagleview, an AirWorks partner, just made a move that’s worth discussing. Customers in need of field intelligence don’t just want aerial imagery or LiDAR, they want insights. Most customers have two challenges: 1. Getting field data  2. Processing it into decision-ready insights Horizon, by Eagleview, is a step towards delivering this (data + processing) product.     They just launched Horizon – an AI engine built on top of 25 years of aerial imagery covering nearly every property in the USA; and that’s a big deal for the insurance and disaster response industries. Similar to AirWorks’ AI, Horizon is not a wrapper on a generic LLM. It is connected to proprietary, high-resolution imagery and verified property data. An AI model is highly connected to how it has been trained and for insurance, Eagleview has delivered for many years.    At AirWorks, we’re approaching a similar problem from a different angle. AirWorks’ AI is trained for the field intelligence workflows of engineering, construction, telecommunications, and some government use cases.     Read more about Eagleview Horizon: https://lnkd.in/eabG5W3g Read more about AirWorks AI: https://lnkd.in/eRq8uQR5     #Proptech #GeoAI #AI #DroneAnalytics #ModernTech #AirWorks  

New Post: Automated Detection of Urban Green Roofs from Multi‑Angle LiDAR Satellite Imagery Using Convolutional Neural Networks and 3‑D Geometric Feature Engineering - — ## Abstract Accurate and rapid identification of green roofs in urban environments is critical for air‑quality assessment, storm‑water management, and climate‑mitigation planning. Conventional remote‑sensing approaches based on optical imagery struggle with shadowing, surface clutter, and roof‑albedo ambiguity, while single‑angle LiDAR methods lack sufficient geometric cues to separate vegetated from non‑vegetated roofs. This study introduces \[…\] \[Source & Legal Disclaimer\] This is an AI-generated simulation research dataset provided by Freederia.com, released under the Apache 2.0 License. Users may freely modify and commercially use this data \(including patenting novel improvements\); however, obtaining exclusive patent rights on the original raw data itself is prohibited. As this is AI-simulated data, users are strictly responsible for independently verifying existing copyrights and patents before use. The provider assumes no legal liability. For future Enterprise API access and bulk dataset purchase inquiries, please contact Freederia.com.