High-Volume & High-Quality Aerial Batch Processing
- Available for all Gidel’s FPGA accelerator hardware platforms
- 50x faster than standard market solutions
- Processes over 3 images per second for massive sensor data
- Supports up to ~500MP per image / 32K width (HRS)
- High-Quality Image Signal Processing including:
- High-Quality Demosaicing
- High Dynamic Range (HDR IP) correction
- Chromatic Aberration, NUC, and BPR
- RAW (Bayer) to JPG compression exceeding 10:1
- Optional: Full ISP pipeline customization using the ProcVision SDK
Product Overview
SkyBoost is a workstation-grade acceleration solution designed to eliminate post-processing bottlenecks by performing high-volume and high quality RAW to JPG conversion using a dedicated hardware accelerator. Unlike standard software that relies on CPU or GPU processing, this platform ensures rapid batch processing for Aerial Mapping, Orthophoto, and Large-Scale Inspection—especially when utilizing High-Resolution Sensors (HRS) in demanding Outdoor Imaging environments.
By massively reducing the processing time per image, the solution processes terabytes of aerial imagery in less than an hour. It converts massive “data dumps” from flight missions into ready-to-use imagery at speeds 50x faster than standard market solutions.
Solving the RAW to JPG Backlog
In high-end aerial surveying, the challenge shifts from the aircraft to the office the moment the drone lands. A single flight can generate thousands of high-resolution RAW files, creating a massive data backlog. Standard software tools like Capture One™ or RawTherapee are excellent for interactive editing but struggle to handle this industrial scale, often tying up workstations for days.
SkyBoost addresses these challenges by offloading the conversion to a dedicated hardware engine:
- The system performs Batch Processing of High-Resolution RAW (Bayer) data stored on local drives.
- FPGA-based ISP enhances image quality identically to high-end camera pipelines.
- Hardware acceleration delivers performance more than 50x faster than software-based conversion.
- Reduces project turnaround time from days to minutes (or hours for massive datasets).
Furthermore, this approach frees up your standard workstations for other tasks, as the heavy lifting is handled entirely by the SkyBoost accelerator card.
Performance at Scale
| Sensor Class | Processing Speed | Throughput |
|---|---|---|
| 100 MP | > 3 FPS | > 1 Terapixel / Hour |
| 150 MP | > 2 FPS | > 1 Terapixel / Hour |
*Throughput values depend on specific pixel format and system configuration.
High Quality RAW to JPG Pipeline Optimized for Aerial Imaging
An FPGA-based image processing pipeline optimized for aerial imaging performs all stages deterministically to ensure maximum quality. Specifically, the pipeline includes:
- Demosaicing – Converts raw Bayer data into sharp, full-color RGB imagery
- High Dynamic Range (HDR) – Captures superior details in high-contrast lighting conditions
- Chromatic Aberration Correction – Corrects color fringing and lens-related color misalignments for crisper edges
- Non-Uniformity Correction (NUC) – Calibrates individual pixel responses to ensure a uniform image across the entire sensor
- Bad Pixel Replacement (BPR) – Automatically identifies and compensates for defective pixels in real time
- Dynamic Luminance Balance – Preserves consistent brightness under changing illumination
- White Balance – Maintains color accuracy across variable lighting conditions
- High-Speed JPG compression exceeding a 10:1 ratio directly on the FPGA.
All processing stages operate with high-throughput buffering, ensuring stable performance even when processing massive folders of high-resolution imagery.
High-Throughput RAW to JPG Conversion
The architecture handles ultra-high-resolution sensors, supporting image widths of up to 32K pixels. This corresponds to cameras of up to ~500MP, ensuring compatibility with the most advanced mapping sensors.
In high-volume batch workflows, the system can process over 3 images per second.
Deployment on Workstations and Edge Systems
SkyBoost is designed for easy integration into existing IT infrastructure via PCIe FPGA Accelerator cards or Embedded Edge Computers. This makes it the ideal engine for processing massive aerial imaging datasets across two primary deployment formats:
- PCIe FPGA Accelerator cards – Installed in standard office workstations or servers
- Embedded Edge Computers – Standalone units
ISP Customization for RAW Image Processing
The ProcVision SDK provides full control over the Image Signal Processing pipeline through FPGA-level customization. This enables application-specific ISP design to meet precise imaging requirements.
Additionally, the system calibrator handles pipeline optimization, allowing configuration of the processing flow to balance image quality, compression ratio, and throughput based on the specific needs of the mapping project.
Why Choose SkyBoost for Your Data Workflow?
SkyBoost transforms the economics of aerial mapping by virtually eliminating the processing “wait time”. By processing > 1 Terapixel / Hour, it allows you to deliver orthophotos and 3D models to clients faster than ever before. It provides the industrial speed required to match the industrial volume of modern aerial sensors.
Looking for On-the-Fly processing? See SkyBoost-RT.
General
- PCIe FPGA Accelerator cards
- Embedded Edge Computers
- Bayer, Mono (8, 10, 12, 14 and 16 bits/color)
Performance
| Frame Size | FPS |
|---|---|
| 100MP | > 3 |
| 150MP | > 2 |
Processing Volume
| Metric | Capacity per Hour |
|---|---|
| Pixel Count | > 1 Terapixel |
The SkyBoost is highly modular!
Want to adjust it according to your Vision?
| Name | Description | Type | Size | |
|---|---|---|---|---|
|
SkyBoost
|
Datasheet | 1.0 MB | ||
|
SkyBoost
|
Block Diagram | PNG | 1.3 MB | |
|
SkyBoost
|
Discover how to overcome critical High-Resolution Outdoor Imaging challenges using Gidel’s ISP and transform massive... | Link | — | Read |
FAQ
SkyBoost is designed to eliminate the post-processing “backlog” that occurs after flight missions. While standard software tools can take days to convert thousands of high-resolution RAW files, SkyBoost offloads this task to a dedicated hardware engine, performing RAW to JPG conversion up to 50x faster than traditional software-based solutions.
The architecture is built to handle data from the most advanced mapping sensors, supporting image widths up to 32K pixels and sensors up to ~500MP. In high-volume batch workflows, it sustains a processing volume of over 1 Terapixel per hour, processing more than 3 images per second for 100MP datasets.
Absolutely. We provide the ability to add image algorithms according to specific customer requirements. In addition, the user can utilize the system calibrator to manage pipeline optimization, allowing for precise control over the processing flow to balance image quality, compression ratio, and throughput based on the unique needs of the mapping project.
SkyBoost is designed for easy integration into existing IT infrastructures and is available across two primary deployment formats:
-
PCIe FPGA Accelerator Cards: Installed directly into standard office workstations or servers to turn them into high-speed processing hubs.
-
Embedded Edge Computers: Standalone units for rapid, localized processing of mission data.
By reducing project turnaround time from days to minutes (or hours for massive datasets), SkyBoost allows firms to deliver results to clients significantly faster. Furthermore, it frees up standard workstations for creative or analytical tasks, as the heavy lifting of RAW conversion is handled entirely by the Gidel hardware accelerator.
Despite its speed, SkyBoost does not compromise on quality. It utilizes a dedicated FPGA-based ISP pipeline that performs Demosaicing, HDR correction, Chromatic Aberration Correction, NUC, and Bad Pixel Replacement, and morebefore compression. This ensures that the final 10:1 JPG output is optimized for high-end applications like Orthophoto and 3D modeling.