High-Speed Frame Grabber for 127 Megapixel Earth Observation Cameras

The evolution of Earth Observation satellites is driving demand for increasingly powerful imaging payloads. Modern commercial and scientific missions are moving toward very large-format image sensors capable of delivering sub-meter resolution, wide-area imaging, and enhanced spectral information. However, capturing and storing data from these high-performance sensors presents a significant challenge: a single image frame can contain hundreds of megabytes of raw data, requiring an advanced high-speed processing and storage architecture.

To address this challenge, CAVU Aerospace has developed a high-performance Command & Data Handling Unit (CDHU) and space-qualified frame grabber architecture designed to interface with 127 Megapixel-class cameras, providing reliable image capture, buffering, processing, and storage capability for demanding Earth Observation applications.

The architecture is based on a Microchip PolarFire® SoC FPGA platform, combining programmable FPGA fabric for deterministic high-speed data acquisition with an embedded processor subsystem for satellite command, control, telemetry and payload management.

Handling Gigabit-Scale Camera Data in Space

High-resolution imaging sensors generate extremely large data streams. For a 127 Megapixel sensor operating with a 10-bit RAW output, each image frame can reach approximately 160 MB, requiring an incoming data rate approaching:

• ~798 MB/s at 5 frames per second
• ~1.6 GB/s at 10 frames per second

Such data rates exceed the capability of conventional satellite computers and require a dedicated payload data path optimized for continuous streaming.

CAVU’s frame grabber architecture solves this challenge through a dedicated FPGA-based acquisition pipeline:

Camera Interface → FPGA Processing Fabric → DDR4 High-Speed Buffer → High-Speed Storage → Satellite Data System

The FPGA fabric performs real-time camera reception, data alignment, packetization and DMA transfer, ensuring that the payload data path remains independent from spacecraft software activities.

FPGA-Based Deterministic Image Acquisition

Unlike processor-only solutions, FPGA-based image acquisition provides:

• Deterministic latency
• Parallel data processing
• High-throughput streaming
• Reliable operation for demanding space environments

The PolarFire SoC architecture uses FPGA fabric as the primary high-speed acquisition engine. Camera data is received through high-speed interfaces, converted into AXI-stream data paths, and transferred into an 8 GB ECC-protected DDR4 ring buffer.

This architecture separates payload data handling from the embedded processor subsystem:

• FPGA Fabric DDR4: dedicated for high-rate image acquisition
• Processor LPDDR4 memory: dedicated for operating system, command handling, telemetry and application software

This separation prevents software activities from interrupting the critical imaging pipeline.