Insight 3
Spatial AI Camera
With its compact size, near-field depth accuracy, and built-in VSLAM, Insight 3 provides stable front-end spatial perception for robotic arms, near-field manipulation devices, and embodied AI data acquisition systems.
Onboard VSLAM
Run real-time visual-inertial localization directly on the camera, enabling robots to track motion and build spatial awareness with less host-side workload.
Dynamic Robustness
Maintain stable spatial perception under vibration, fast motion, sudden obstacles, and changing real-world environments.
Onboard AI Models
Run edge AI perception models directly on the device to deliver real-time semantic understanding while reducing bandwidth and host compute demand.
Holeless Depth Mapping
Generate dense, AI-enhanced depth maps with fewer missing areas, giving robots cleaner 3D structures for navigation, obstacle avoidance, and data capture.
Spec
| Use environment: | Indoor |
|---|---|
| Stereo Image sensor technology | Global Shutter |
| Ideal Range | 0.06m to 5m |
| Baseline: | 35 mm Standard |
| Inertial measurement unit: | 6 Axis/± 24 g |
| Onboard Visual SLAM solution | Visual Inertial Odometry(VIO) |
| Onboard AI Models | Object Detection / Human Pose Estimation / Gesture Estimation, etc |