Key Features:
-
Complete ISP pipeline
Includes black level correction, digital gain, demosaicing, anti-aliasing, auto white balance, auto exposure, gamma correction, and tone mapping.
-
PVA-Accelerated Processing
Executes ISP algorithms on the NVIDIA Programmable Vision Accelerator, freeing CPU and GPU resources. -
Software-Defined ISP Pipeline
Provides a configurable alternative to fixed-function and black-box ISP implementations. -
Customizable Algorithms
Allows ISP stages to be tuned, replaced, reordered, or extended for application-specific image-quality requirements. -
Linear and Companded RAW Support
Supports linear Bayer data and companded HDR RAW input through configurable decompanding. -
Multiple Output Formats
Produces standard formats suitable for downstream processing, including RGB888 and NV12. -
Sensor-Agnostic Design
Works with Bayer RAW sensors independently of the camera manufacturer. -
Runtime Configuration and Tuning
ISP parameters can be adjusted without redesigning a fixed hardware pipeline. -
Low-Power Image Processing
Uses dedicated vision hardware to reduce reliance on power-intensive GPU processing. -
Real-Time Embedded Integration
Designed for low-latency camera pipelines on supported NVIDIA Jetson platforms equipped with PVA. -
Source-Code Availability
Enables customers to inspect, tune, replace, and add algorithms when purchased with source-code access.
Example Use Cases
- Multi-camera perception platforms: Offloads ISP processing from the GPU, preserving resources for perception and AI workloads across multiple camera streams.
- Custom processing pipelines for machine perception: Enables algorithms to be tuned for feature preservation and model accuracy rather than visual appearance.
- Fully controlled imaging pipelines for safety-critical applications: Provides visibility and control over each processing stage, supporting deterministic behavior, validation, and traceability.
- Offline RAW image processing: Processes previously captured Bayer RAW data for tuning, testing, dataset preparation, and algorithm validation.
- Low-power edge AI camera systems: Uses the dedicated PVA hardware to reduce GPU usage and improve overall power efficiency.
-
High-resolution camera systems: Accelerates computationally intensive ISP operations to support high-resolution image processing with low latency.
Supported Platforms
- NVIDIA Jetson Thor.
- NVIDIA Jetson AGX Orin.
- NVIDIA Jetson Orin NX.
Platform compatibility is determined by the availability of supported NVIDIA PVA hardware.
What’s Included
- C++ library: Core PVA-accelerated ISP library with APIs for configuring, integrating, and controlling the image-processing pipeline.
- Example applications: Ready-to-use examples demonstrating pipeline configuration, image processing, and integration into custom applications.
- GStreamer element: A GStreamer plugin that simplifies integration and deployment within existing multimedia and camera pipelines.
- Debugging and calibration tool: A visual tool for inspecting intermediate results, adjusting ISP parameters, and tuning image quality during development.
Why RidgeRun PVA ISP?
- Backed by RidgeRun expertise: Built by a team with extensive experience in GStreamer, embedded Linux, camera systems, and hardware-accelerated multimedia.
- Free GPU resources: Offload image processing to the PVA, preserving GPU capacity for AI and perception workloads.
- Gain full pipeline control: Inspect, configure, tune, replace, and extend ISP stages without depending on a black-box implementation.
- Reduce power consumption: Use dedicated PVA hardware for efficient processing in power-constrained edge systems.
-
Build application-specific ISP workflows: Adapt the pipeline to your sensor, image-quality, HDR, or machine-perception requirements.
Learn more in our developer's wiki.
Monthly Subscription
Hire RidgeRun at a flat monthly rate for a maximum of 8 hours per day, 40 hours per week each month. Get fast interaction with engineers working on your project, weekly 45 minute calls to update you on our progress, and the complete engagement and expertise of our engineering team.
Work With Us