Meet us September 15–17 at two trade shows: IFTM 2026 and NRF 2026.

Products  ›  Applied computer vision

The most advanced spatial computing tools, ready to use.

Wemap Vision Engine turns low-cost 360° images into analysed, navigable and augmented spaces. From immersive virtual tours to automatic asset detection, AR navigation and visual search.

Engine Wemap Vision
R&D 3 PhDs · 10 Engineering
Deployment > 100 sites / month
See case studies → Request a technical demo

CAPABILITIES

Ship multiple high value services, in one simple move.

Ultra-fast 360 video capture and generation of a visual model of your space, laced with AI tools.

Advanced virtual tour

Georeferenced 360° viewer — explore the space as if you were there. Localised annotations, measurements, deep links to every point and orientation in space. Continuous update: rapid recapture without heavy redeployment.

Lightweight 360 ViewerFully ImmersiveDeepLink

Clients

Galeries Lafayette · Château de Versailles · RATP

60,000 m²

explorable — Château de Versailles

Search, detection and identification

Visual search engine powered by a Vision Language Model — find equipment, a brand, a product or an area by text description or image capture. Automatic detection and inventory of equipment and signage in 360° images, trainable on your business objects (YOLO, DINO, Segment Anything) — up to 85% savings on field reference dataset creation.

VLMAssetDetectionAutoInventory

Clients

Galeries Lafayette · Leading North American Airport · RATP

< 3 seconds

to locate an object or a trace — without prior knowledge nor annotation

Vision-based positioning

Computer vision-based positioning — the smartphone locates itself by comparing its camera feed to the space's 360° model, with no beacon or infrastructure. Precise blue dot to 1-3 metres, acquisition in under 5 seconds.

VPSZeroInfrastructureBlueDot

Clients

SNCF (40+ gares) · Hôpital Sainte Marie · Westfield

1–3 m

accuracy, with no infrastructure to deploy

Augmented reality

Step-by-step AR navigation (floor chevrons overlaid on reality, full-scale instructions) and visual AR search (POIs visible on raising the phone). Works web (JS) and native (iOS, Android, Flutter). Offline available.

ARNavigationVisualSearchARWebAROfflineAR

Clients

Keolis Nancy · ADP · Ville de Colmar

98%

of users find AR useful — Keolis Nancy

CASE STUDIES

Immediate operational gains.

Galeries Lafayette Retail

AI virtual viewer — VLM, 3D annotations, measurements, visual search engine. Explore the store environment and locate any asset in an instant, remotely.

< 3 seconds

to create an inventory of assets

Oxysign Industry · Transportation

Inbound and outbound inspections at railway technical centers: instantly create an explorable and searchable record of actual conditions.

15 minutes

to capture a full trainset

See all Vision case studies →

OPEN BY DESIGN

Your AI models. Your data. Your formats.

Wemap Vision Engine is trainable on your business objects and compatible with major computer vision frameworks.

Train your models on our datasets

The 360° images captured during deployment form a structured, geolocated dataset — usable to train any computer vision algorithm.

YOLODINOSegment AnythingGaussian SplattingPyTorchONNX

"Your R&D team keeps control of your models. Wemap provides the data, the capture pipeline, and the spatial and segmentation layer."

Your data, your formats

Full portability, no proprietary lock-in, open formats throughout the pipeline.

  • ✓ GeoJSON, CSV, standard format exports
    Your data is exportable at any time, in interoperable formats.
  • ✓ BIM compatibility (IFC, Revit)
    Integration into your BIM workflows and existing space repositories.
  • ✓ On-premise available
    Deploy Wemap Vision Engine in your own cloud or on-site infrastructure.
  • ✓ No GAFAM cloud dependency
    Technology designed in France and operated in Europe, deployable on any major cloud provider.

Ready to explore your spaces with Wemap Vision?

We help you test Wemap Vision in just a few days: you get a location-enabled site featuring your own data, allowing you to envision the services of tomorrow.

Request a technical demo → See the Platform
FAQ

Frequently asked questions

Wemap Vision works with any consumer 360° camera (Insta360, Ricoh Theta). A camera costing a few hundred euros is sufficient. No specialist equipment (LIDAR, industrial scanner) is required.
Capture takes a few hours with a 360° camera. Processing and reference generation are automated. Modified areas can be partially recaptured for a quick update.
The VLM engine lets you search for an object or equipment by text description across the 360° reference. Models can be trained on your specific business objects (YOLO, DINO, Segment Anything).
Yes. Wemap offers private EU cloud or on-premise hosting. 360° images are processed and stored in your environment. No data passes through third-party services. Native GDPR compliance.
Yes. Wemap uses open standard formats (GeoJSON, point cloud, JSON annotations). Data can be exported to your CMMS, BIM or ERP tools. Full reversibility guaranteed.