Bird’s Eye View for NVIDIA Jetson: Real-Time 360° Surround-View
 |
| Getting Started |
|---|
| User Guide |
| Calibration Guide |
| Calibration Guide Legacy |
| GStreamer |
| Performance |
| Performance Measurements |
| Contact Us |
To get started with BEV, follow the Quickstart Guide.
What Is RidgeRun Bird’s Eye View?
RidgeRun’s Bird’s Eye View (BEV) is a real-time multi-camera stitching software system that transforms images from multiple cameras into a seamless 360° top-down (aerial) view. Designed for embedded platforms such as NVIDIA Jetson and NXP i.MX8, BEV uses on-board hardware accelerators to achieve low-latency image warping, blending, and rendering. Its portable architecture allows it to be integrated into a wide range of robotics, automotive, and industrial vision applications.
Who is it for?
RidgeRun’s Bird’s Eye View is used by teams building:
- Automotive surround-view and parking-assist systems
- Robotics navigation and autonomous mobility
- Safety, monitoring, and situational-awareness applications
- Industrial vehicles, forklifts, and AGVs
Why Bird’s Eye View Matters?
Robots, vehicles, and industrial systems operate in environments full of blind spots and unpredictable motion. A real-time bird’s-eye view eliminates these blind zones by combining multiple cameras into a single, intuitive 360° perspective. This improves navigation, enhances safety, and enables faster, more reliable decision-making, especially in applications where milliseconds matter.
With RidgeRun’s optimized pipelines for Jetson and i.MX8, teams achieve high-quality situational awareness with low latency and minimal compute overhead.
How does Bird’s Eye View work?
The Bird’s Eye View (BEV) software transforms images from multiple perspective cameras into a unified 360-degree top-down view using advanced perspective-transformation and multi-camera stitching algorithms. Raw camera feeds are captured by onboard sensors and processed through RidgeRun’s hardware-accelerated image transformation pipeline running on embedded platforms such as NVIDIA Jetson and NXP i.MX8. A JSON configuration file defines camera models, calibration parameters, and layout geometry to ensure precise alignment and seamless blending between views. The result is a real-time, low-latency bird’s-eye visualization ideal for robotics navigation, automotive surround-view systems, industrial vehicles, and other embedded AI applications requiring reliable 360° situational awareness.
Key Features
- Surround 360 Video
- Integration with GStreamer
- Portable, currently supported in NVIDIA Jetson and NXP i.MX8
Performance
RidgeRun BEV achieves real-time stitching at 30–60 FPS on NVIDIA Jetson and i.MX8 platforms. Benchmarks for Jetson include:
- 6 camera top-down view at 1080p @ 60 FPS
- BEV processing time for 6 cameras at 1920x1080 as low as 14 ms
- GPU utilization typically below 65% on Jetson Orin.
See our Performance Measurements for NVIDIA Jetson page for more information.
Try RidgeRun Bird’s Eye View
RidgeRun offers free evaluation binaries of Bird’s Eye View so you can validate 360° surround-view on your own hardware before committing to a full engagement. To get started, send us an email at support@ridgerun.com with a brief description of your project, including:
- Target platform (e.g., Jetson Orin, Xavier, Nano, NXP i.MX8, PC, other)
- Number and type of cameras (FOV, resolution, framerate, interfaces)
- Desired output (top-down view, panorama, 360° surround-view, etc.)
- Performance goals (target resolution and FPS, latency constraints)
Our engineering team will follow up with tailored guidance, evaluation binaries, and next steps to help you quickly determine if RidgeRun’s Bird’s Eye View is the right fit for your application.
Why choose RidgeRun?
- Built for embedded hardware — optimized for Jetson & i.MX8 with CUDA, NVMM, and OpenGL acceleration
- Real-time performance — low-latency stitching and top-down rendering at 30–60 FPS
- Flexible camera support — fisheye, wide-FOV, RGB, mono, multi-camera rigs
- Proven in real deployments — validated in robotics, industrial vehicles, and safety applications
- Expert engineering support — direct access to the developers who maintain and optimize BEV
- Comprehensive documentation — includes detailed Quickstart and Calibration guides for streamlined integration
- Source code access — obtain full access to the BEV library source code for customization, evaluation, and integration
FAQ
What is RidgeRun’s Bird’s Eye View (BEV)?
BEV is a real-time 360° surround-view software solution that stitches multiple camera feeds into a seamless top-down view on embedded platforms like NVIDIA Jetson and NXP i.MX8.
Which platforms does BEV support?
BEV runs on NVIDIA Jetson (Orin, Xavier) and NXP i.MX8 platforms, using CUDA, OpenGL, and hardware accelerators for real-time performance.
How many cameras does BEV support?
Typical deployments use 4–6 cameras, but BEV can be customized to support additional sensors depending on platform resources.
What cameras resolutions are recommended for BEV?
BEV has been evaluated with cameras up to 4K resolution.
Does BEV work with fisheye or wide-angle lenses?
Yes — RidgeRun offers add-ons for customers that require calibration and distortion-correction tools for fisheye, wide-FOV, and other non-linear lenses.
What latency can BEV achieve?
Depending on platform and resolution, BEV can operate in real time (e.g., 30–60 FPS) with low-latency GPU-accelerated processing.
How do I evaluate BEV?
Contact support@ridgerun.com with your use case to receive free evaluation binaries and integration guidance.