SkyBoost Accelerates Aerial Photogrammetry Post-Processing
Gidel, a technology leader in FPGA-based imaging and vision solutions, today announced the launch of SkyBoost, an FPGA hardware accelerator designed to accelerate aerial photogrammetry post-processing. By offloading computationally intensive RAW to JPG conversion to a dedicated FPGA hardware accelerator, SkyBoost helps aerial mapping teams transform large post-flight datasets into analysis-ready aerial imagery far faster than software-only workflows.
Solving the Industrial-Scale RAW to JPG Backlog
In high-end surveying, the bottleneck often shifts from the aircraft to the office as soon as a mission ends. A single flight can generate thousands of high-resolution RAW files, creating a backlog that ties up workstations and delays deliverables. SkyBoost addresses this by moving conversion and image processing into dedicated FPGA hardware, so post-processing is no longer the weakest link in the workflow.
Accelerating High-Volume Aerial Photogrammetry and Drone Mapping Workflows
SkyBoost is engineered for the throughput requirements of modern high-resolution sensors as drone mapping and UAV mapping projects scale toward higher resolutions and larger datasets. In batch mode, SkyBoost performs high-volume processing of high-resolution RAW (Bayer) data directly from local storage and delivers total throughput of over 1 Terapixel per hour. This enables organizations to process terabytes of aerial imagery in less than an hour, maintaining speeds of over 3 images per second for 100MP sensors and over 2 images per second for 150MP sensors.
Advanced FPGA ISP Pipeline for Precision Orthophoto and 3D Photogrammetry
Quality is not sacrificed for speed. SkyBoost uses an FPGA-based Image Signal Processing (ISP) pipeline that runs deterministically to support consistent results for orthophoto and 3D photogrammetry workflows, especially in outdoor imaging conditions with challenging lighting and high dynamic range scenes. The pipeline includes high-quality demosaicing to convert Bayer data into sharp RGB imagery, HDR correction for high-contrast lighting, and chromatic aberration correction to reduce color fringing. It also includes Non-Uniformity Correction (NUC) and Bad Pixel Replacement (BPR) to help maintain uniform image quality across the sensor. SkyBoost supports ultra-high-resolution sensors up to approximately 500MP, including image widths up to 32K pixels.
Why FPGA for Aerial Photogrammetry?
Unlike CPU-based image processing, SkyBoost performs deterministic FPGA image processing with predictable throughput, enabling organizations to process extremely large aerial imaging datasets while maintaining consistent image quality and freeing workstation resources for other tasks. SkyBoost is built entirely on Gidel’s in-house FPGA platform and can be customized using the ProcVision SDK or expanded with additional FPGA image processing IPs to meet specific application requirements.
Why Choose SkyBoost for Aerial Photogrammetry?
- Accelerate project turnaround times from days to minutes by processing over 1 Terapixel per hour.
- Convert RAW Bayer data to high-quality JPG images up to 50x faster than software-only workflows.
- Offload computationally intensive image processing to dedicated FPGA hardware.
- Maintain industrial-scale throughput for demanding datasets, including support for sensors up to ~500MP and 32K pixel widths.
- Enhance image quality using a deterministic FPGA-based ISP pipeline that includes white balance, dynamic luminance balance, NUC, BPR, HDR, and chromatic aberration correction.
- Customize the ISP pipeline using ProcVision SDK to meet specific project imaging requirements.
Workflow Fit
SkyBoost integrates into the post-capture phase of professional mapping missions:
Data Capture: Teams capture high-resolution RAW or mono imagery during the flight mission.
Data Transfer: Teams copy datasets from aircraft media to local or server storage.
Batch Ingest: The user selects a folder of RAW files for batch conversion of RAW images into high-quality JPEG output.
Processing Profile: The user selects the processing preset and output settings, for example demosaicing, white balance, dynamic luminance balance, and JPEG quality.
Hardware Offloading: SkyBoost performs the processing using its dedicated FPGA processing engine rather than the host CPU.
Deterministic ISP: The hardware applies the selected correction pipeline, including image normalization, based on configuration.
High-Speed Compression: The engine performs JPEG compression with ratios exceeding 10:1, depending on scene content and quality settings.
Final Output: Processed images are saved for immediate use in orthophoto and 3D photogrammetry software.
Target Applications for Aerial Photogrammetry
- Aerial mapping
- Orthophoto production
- Ultra-high-resolution aerial imaging
- 3D modeling and 3D photogrammetry
- Smart cities
- Aerial imaging
- Post-disaster assessment
- Large-scale inspection
- Environmental monitoring
- Precision agriculture
- Drone mapping and UAV mapping
Watch SkyBoost in Action
See how SkyBoost accelerates aerial photogrammetry post-processing by converting high-resolution RAW imagery into JPEG outputs through a dedicated FPGA-based processing pipeline.
This exhibition demo shows the image-processing flow from RAW input to JPEG output, with the RAW images displayed on the left and the generated JPEG results displayed on the right. The demo illustrates accelerated RAW image processing and JPEG generation for mapping and photogrammetry applications.
Note: Performance figures are based on Gidel internal benchmarking in batch mode under controlled conditions. Actual results may vary depending on dataset characteristics, selected ISP options, storage bandwidth, and host system configuration.
Deployment Flexibility
SkyBoost is available as a PCIe FPGA accelerator for Windows and Linux workstations and servers. The same FPGA image processing pipeline can also be deployed on Gidel’s FantoVision system when a compact, stand-alone embedded solution is required.
Accelerating the Future of Aerial Photogrammetry
As aerial imaging sensors continue to increase in resolution, post-flight processing is becoming a critical bottleneck for mapping and photogrammetry organizations. SkyBoost addresses this challenge by combining deterministic FPGA image processing with high-throughput JPEG compression, significantly reducing project turnaround times while maintaining consistent image quality across large aerial datasets.
Whether processing hundreds or thousands of high-resolution aerial images, SkyBoost enables surveying, mapping, and geospatial professionals to move from data capture to analysis faster. Its scalable FPGA architecture also helps organizations keep pace with growing dataset sizes, higher sensor resolutions, and increasingly demanding photogrammetry workflows.
👉 Explore Gidel SkyBoost
👉 Schedule a Demo
👉Visit Geo Week for Industry News
