Design VDEgo - Egocentric Video Data Collection Service at Scale | FPV Head-Mounted Camera Planning, Solution
Product Description
Say goodbye to bulky rigs and tangled wires. VDEgo features an ergonomic, lightweight head-mounted design that allows operators to move naturally. Capture authentic data in factories, warehouses, retail stores, or domestic environments without restricting human movement.
Streamline your data processing pipeline. VDEgo supports independent data collection based on specific semantic tasks. Every task you record corresponds to a complete, neatly organized data package, making it incredibly easy to label, sort, and utilize for machine learning models.
VDEgo doesn't just record video; it captures a complete sensory footprint. The device outputs:
- High-Resolution Color Video: MP4 / H.265 format.
- Image Timestamps: For frame-perfect synchronization.
- IMU Data: High-frequency inertial measurements (with timestamps).
- Audio Recording: Capturing environmental soundscapes.
- Calibration Parameters: Ensuring immediate usability for spatial computing.
To further accelerate your R&D, VDEgo comes equipped with a proprietary offline trajectory calculation tool, allowing your engineering team to extract precise spatial movement paths directly from the collected data.
We understand that data collection at scale needs to be both effortless for the operator and secure for the enterprise. The VDEgo workflow is designed for maximum efficiency:
Operators simply wear the VDEgo headset and press a single physical button on the device to start or stop recording. No complex interfaces or tethered computers are required during active capture. Data is safely stored on a local SD card during the session.
Once the daily data collection is complete, VDEgo offers two highly flexible upload pathways to suit your security and infrastructure needs:
- Cloud Upload: The device connects to Wi-Fi (easily configured via a quick WeChat QR code scan) and automatically uploads the day's data to the cloud.
- Secure Localized Deployment (LAN): For data factories, partner manufacturing plants, or academic institutions with strict data privacy protocols, VDEgo supports localized server deployment. Data is transmitted directly to your internal local servers via your private Wi-Fi LAN—completely bypassing the public internet to guarantee absolute data security and confidentiality.
Features a dual-camera setup perfect for standard stereoscopic vision data collection. Ideal for retail, home services, and general AI training.
Features a four-camera array for an ultra-wide field of view and enhanced spatial depth perception. Engineered for complex industrial assembly, special operations, and advanced robotics R&D.
To ensure data integrity and seamless integration into your AI training pipelines, VDEgo exports all collected data in a custom compressed package. Virdyn provides a dedicated Decompression API to help developers easily extract and parse the data.
Each data package contains comprehensive, perfectly synchronized multimodal data:
- Video Data: High-quality color video recordings (MP4/H.265).
- IMU Data: High-frequency motion tracking data.
- Audio Data: Synchronized environmental sound recordings.
- Precise Timestamps: Frame-level hardware timestamps for perfect vision-motion synchronization.
- Calibration Parameters: Built-in intrinsic and extrinsic parameters for both the cameras and the IMU, ready for immediate spatial calculation.
VDEgo is the premier data collection tool for a wide variety of industries:
- Industrial Assembly & Manufacturing: Capture expert technician workflows to train automated robotic arms.
- Warehouse & Logistics Sorting: Record first-person picking and packing data for AI optimization.
- Home & Domestic Services: Collect natural human interaction data for domestic companion robots.
- Retail & Merchandising: Analyze first-person shopper behavior and shelf-stocking procedures.
- Special Operations: Document complex, hands-free tasks in hazardous or specialized environments.
VDEgo adapts to your operational environment with two intuitive and highly efficient data collection modes, ensuring you can capture data effortlessly in any scenario:
- How it works: Simply press the physical button on the VDEgo headset to start or stop recording instantly.
- Key Feature: Supports simultaneous, real-time data uploading to your designated server, eliminating the need for manual data transfer and keeping your hands completely free.
- How it works: Connect your PC or smartphone to the same Wi-Fi network as the VDEgo device and access the built-in web dashboard. No app installation required!
- Key Features:
- Remote Control: Start and stop data collection remotely.
- Live Preview: Monitor the first-person camera feed in real-time.
- Data Management: Download recorded files directly to your local device or push them to a designated server.
- Status Monitoring: Easily check local storage capacity and track real-time upload progress.
VDEgo delivers frame-synchronized multimodal datasets combining high-resolution H.265 video, precise IMU movement telemetry, audio, and factory calibration parameters, providing complete ground-truth information required for robotic imitation learning.
VDEgo-C2 features a binocular dual-camera setup ideal for standard stereoscopic data capture. VDEgo-C4 utilizes a quad-camera system offering an ultra-wide field of view and enhanced spatial depth perception for complex industrial workflows.
Yes. Data is continuously written to an internal SD card during recording. For secure facilities, VDEgo supports complete localized LAN server deployment, completely bypassing the public internet to protect confidential proprietary operations.
No app installation is needed. Operators can trigger recordings directly with the physical on-device button or connect to the built-in wireless web dashboard via any mobile phone, tablet, or PC browser on the network.
Recorded sessions are packaged into organized task bundles. Engineering teams can use the included Decompression API and offline trajectory computation tool to readily extract synchronized video, sensor data, and coordinate tracks.
