EL-iVISION
The spatial-intelligence platform behind everything we deliver — from multi-sensor observations to actionable spatial intelligence.
EL-iVISION is a localization-first platform. Positioning is established before anything is mapped, and mapping before anything is interpreted, because a perception result anchored to a wrong pose is worse than no result at all. The stages below describe how data moves through the platform.
Platform Architecture
Four stages, each with a defined responsibility and a defined interface to the next.
- 01
Input Sensors
- 3D LiDAR
- RGB Camera
- Thermal Camera
- IMU
- RTK-GNSS
- 02
Spatial Core
- Time Synchronization
- Calibration
- Odometry
- SLAM / Localization
- Sensor Fusion
- 03
Intelligence Layer
- 3D Perception
- Object Tracking
- Semantic Understanding
- 04
Outputs
- Pose
- Trajectory
- 3D map
- Object states
- Infrastructure information
- Operational analytics
Every EL-iVISION capability in one map, from sensor input through to autonomous operation.
Input Sensors
EL-iVISION is designed around the sensors already present on your platform. No proprietary sensor purchase is required.
- 013D LiDAR
- Geometric structure and range. The primary source of metric scale.
- 02RGB Camera
- Appearance, texture and semantics for classification and recognition.
- 03Thermal Camera
- Radiometric signature — vital where illumination fails or is misleading.
- 04IMU
- High-rate inertial measurement that bridges gaps between slower sensors.
- 05RTK-GNSS
- Global reference where sky visibility allows, for georeferenced output.
Spatial Core
The part that establishes geometry. Everything else in the platform depends on it being correct.
- 01Time Synchronization
- Aligning streams that arrive at different rates and with different latencies into one consistent timeline.
- 02Calibration
- Intrinsic and extrinsic parameters, monitored rather than assumed constant across an operating life.
- 03Odometry
- Incremental motion estimation from LiDAR, vision and inertial measurement.
- 04SLAM / Localization
- Pose-graph estimation with loop closure, plus localization against an existing map.
- 05Sensor Fusion
- Combining modalities so that the failure of one does not become the failure of the system.
Intelligence Layer
Where established geometry acquires meaning.
- 013D Perception
- Detection and segmentation in three dimensions, with real-world extents rather than image-plane boxes.
- 02Object Tracking
- Temporal association and state estimation, so an object keeps its identity across frames.
- 03Semantic Understanding
- Classifying scene elements — lanes, signs, surfaces, assets — and anchoring them spatially.
Outputs
Machine-readable results, delivered through interfaces your team already uses.
- Pose
- Trajectory
- 3D map
- Object states
- Infrastructure information
- Operational analytics
Deployment Model
How EL-iVISION is delivered into an existing engineering programme.
Onboard / Edge
Runs on the compute already on the platform, for closed-loop use where latency matters.
Offline / Batch
Post-processed mapping and analysis over recorded sessions, for survey and inspection work.
Hybrid
Onboard localization with offline map construction and refinement between operating sessions.
Integration is through ROS 2 topics and services, or a native C++ library for teams outside the ROS ecosystem.
Discuss How EL-iVISION Fits Your Stack
Start with a technical conversation about your platform, sensors and constraints.