TERRA AI
See in deep detail. Gain deep understanding.
Advanced perception to meet your mission
To operate on a farm, or a construction site or even a warehouse, a robot needs more than sight—it needs a semantic understanding that peels back the layers of an environment to reveal the intricacies that truly matter.
TERRA AI reveals a level of deep detail, with a neural network that translates ZED cameras texture and color data into nuanced insights that outshine alternatives like LiDAR or radar.
Sense, plan and act, without interruption.
TERRA AI is a groundbreaking large neural network architecture that is able to deliver an unmatched level of understanding to off road autonomous robots.
It provides a comprehensive understanding of an environment by fusing:
- Depth Estimation for spatial awareness
- Visual-Localization and GNSS fusion to determine a robot’s position
- Semantic understanding for obstacle classification
- Robust detection and tracking of objects and people
Ready to deploy, TERRA AI’s general-purpose capabilities are run through low-power embedded computers that can be integrated into any fleet.
Unlock vision-driven precision
TERRA AI is vision-based, using images captured by monocular and stereo ZED cameras as its primary input. Significantly cheaper than LiDAR or radar solutions, these cameras can be strategically positioned to provide a panoramic 360° view around any robot.
These images then undergo self-calibration to determine all camera positions, multi-camera fusion, and time-fusion, to output critical data for robots autonomy: depth maps, localization data, semantic segmentation, object and people detection and tracking for safety.
Trained with TERRA SIM, our engine for building AI vision for robotics
Our TERRA SIM delivers:
- Accurate sensor simulation with photorealistic rendering
- Procedural generation of infinite worlds
- Populated with AI agents who seek collision
- Diversity of scenes, weather, challenging conditions, vehicles
- Resolution of long-tail issues
Integrate TERRA AI into your fleet
We provide plug-in advanced perception to fleets of all sizes. ZED cameras capture rich data in texture and color; the ZED Box then processes this data before transmitting it to TERRA AI’s neural network.
Tried and tested technology
22,000+
BUSINESSES
100,000+
DEVELOPERS
13+
YEARS OF EXPERIENCE
50,000+
CAMERAS IN ACTON
Experience the benefits of unmatched understanding
Our plug-in AI perception unlocks transformative mission potential in every system.
A scalable solution for any application
TERRA AI is designed to be effective in indoor and outdoor environments, giving robots of any size persistent vision through occlusions.
Advanced perception, at a fraction of the price
TERRA AI is a vision-based system, using cameras to replicate human-like sight. This approach is not only more effective for precision-based tasks than LiDAR is, but is ten times less costly to integrate into your fleet.
Beyond 360 vision
TERRA synthesizes 3D perception, 4D spatiotemporal fusion and multi-view perception. This gives robots the ability to safely navigate any environment, without blind spot
Efficient memory and compute
TERRA AI is built to scale, and has been designed to manage large computational loads through low-power embedded system.
Ready to deploy
TERRA AI’s general-purpose capabilities can be applied to diverse applications across multiple industries.
Agriculture
Construction
Logistics
A new era of vision-driven precision
Start building exciting new applications that recognize and understand your environment.
