Description
- Slamtec Aurora S 3D AI-VSLAM Spatial Perception System
- Combines vision and AI-VSLAM technology
- Outputs real-time 3D mapping and localization
- Includes fisheye and stereo depth cameras
- Supports over 100 object semantic recognition
- Integrates with VLA/VLN and multiple SDKs
The Slamtec Aurora S 3D AI-VSLAM Spatial Perception System is an advanced AI spatial perception solution from SLAMTEC's Insight Series. It combines vision, inertial navigation, and proprietary AI-VSLAM technology to deliver precise 3D perception, mapping, and semantic understanding for robots and intelligent agents. This system is engineered for challenging environments and applications, providing real-time, high-quality spatial perception outputs without requiring additional computational resources.
The system features built-in computing algorithms that directly output mapping and localization results. It includes a 180° fisheye color camera and a 120° ultra-wide-angle stereo depth vision, enabling pixel-level semantic recognition of over a hundred objects. The multi-modal fusion capability facilitates semantic mapping, while the one-click dense 3D reconstruction supports 3DGS solutions. The Aurora S seamlessly integrates with VLA/VLN training systems and is an open platform that supports various SDKs, including C++, ROS1/ROS2, and Python.
Its compact, fully integrated design incorporates real-time loop detection and error correction. The binocular dense depth sensing system delivers wide-angle point clouds at 15 frames per second. The high-performance real-time semantic segmentation and object recognition engine supports the recognition of 18 outdoor scene categories and over 80 indoor object types. The system offers rich interfaces and expandability, supporting a DC 9–24V power supply or USB Type-C PD3.0.
- 1x Slamtec Aurora S 3D AI-VSLAM Spatial Perception System
- 98 x 84.8 x 26 mm (L x W x H)
- Max. Mapping Area: >1,000,000m2
- Relocation: Support entire relocation, accuracy ±5cm
- Map Management: Support mapping continuation, loading and saving
- Mapping & Localization Mode: AI deep-learning VSLAM (primary), optional LiDAR fusion
- Multi-sensor Synchronization: Hardware-level time synchronization
- Camera Specifications: Stereo fisheye, 60mm baseline, FOV 180°, Global Shutter, RGB
- Camera Frame Rate: Typical 15Hz, customisable 10/30Hz
- Dense Depth Camera Capability: End-to-end deep learning; >90% detection rate; works in strong light and weak texture environments
- AI Object Recognition & Segmentation: Support 18 outdoor scenarios + 80 indoor scenarios, customizable upon request
- Power Consumption: 10 W (Typical, excluding LiDAR)
- Operating Temperature: -20°C ~ 50°C
- Startup Temperature: ≥0°C
- Storage Temperature: -20°C ~ 60°C