Powering Physical AI with Real-World Ego Datasets
FPV Head-Mounted Egocentric Camera | Video Data Collection Camera
Version: Full Hat
Quantity
Share
VDEgo-C2 Product Description
Key Features & Performance
True First-Person Perspective (Ego-Vision)
Designed to replicate the human field of view, the VDEgo-C2 captures authentic human-eye perspective data. It is the ideal head mounted camera for data collection involving complex hand-object interactions, tool usage, and spatial navigation.
Synchronized Binocular RGB + IMU Data
Equipped with dual RGB cameras and a high-precision Inertial Measurement Unit (IMU), the device ensures microsecond-level synchronization between visual feeds and motion data. This multimodal synchronization is crucial for reconstructing precise device trajectories and training robust physical AI models.
Ultra-Lightweight & Ergonomic Design
Weighing less than 300g, the VDEgo-C2 features an optimized, balanced headband design. It eliminates neck fatigue during extended sessions, enabling researchers and operators to perform all-day, real-world data collection in comfort.
Secure Local Storage & Offline Capture
To protect proprietary and sensitive training data, the VDEgo-C2 writes data in real-time to an onboard high-speed SD card with hardware-level encryption support. Built-in physical buttons and acoustic feedback (prompt tones) allow for instant, independent offline recording without requiring a continuous network connection.
Wireless Control & Real-Time Visualization
Our integrated video data collection system features dual-terminal Web UI access (PC and mobile). Users can wirelessly preview low-latency live feeds, remotely trigger start/stop recording, configure camera parameters, and monitor synchronous upload progress over local Wi-Fi.
Open APIs & Trajectory Reconstruction
VDEgo-C2 comes standard with a dedicated data parsing toolkit. It supports raw data decryption and offline SLAM/trajectory reconstruction. With fully open APIs and SDKs, developers have unrestricted access to raw sensor data for seamless secondary development.
Technical Specifications
| Parameter | Specification |
|---|---|
| Camera Type | Binocular RGB Camera (Stereo Vision) |
| Sensor Synchronization | Hardware-level synchronization (RGB Video + IMU) |
| IMU | High-precision 6-axis / 9-axis Inertial Measurement Unit |
| Weight | < 300g (Ultra-lightweight ergonomic headband) |
| Storage | Local High-Speed MicroSD Card (Supports hardware encryption) |
| Connectivity | Dual-band Wi-Fi (for Web UI & wireless transfer), USB-C (for charging & data) |
| Control Interface | Physical on-device buttons (with audio prompts) + Wireless Web UI |
| Compatibility | Fully open APIs & SDKs; supports ROS/ROS2 and custom Python pipelines |
| Mounting System | Versatile clamp structure compatible with textile-based headwear/caps |
Primary Applications
- EGO First-Person Multimodal Datasets: Capture synchronized video and motion data to build gold-standard datasets for human behavioral analysis.
- VLA (Vision-Language-Action) Model Training: Provide high-quality bimanual manipulation ego video datasets to help robots learn imitation, manipulation, and action planning.
- World Model Training: Utilize synchronized IMU trajectories and stereo video to reconstruct 3D scene geometry and dynamic environmental physics.
- Academic & Robotic Research: An open-source friendly platform for computer vision departments, university labs, and AI research institutes.

