Unibap LOOM
Unibap Space Solutions AB
Real-time hyperspectral image preprocessing pipeline that enables satellites to autonomously process hyperspectral data in orbit and deliver actionable insights with minimal latency.
Technical specifications
- Output format
- Corrected GEO-TIFF images
- Max bands
- Up to 32 (expandable)
- Pixel width
- 4096 pixels (configurable)
- Processing
- Multi-threaded single binary
- Georeferencing
- UTM coordinate-based affine transformation
- Installation
- .deb packages for Linux systems
- Image corrections
- Band co-registration, statistical de-noising, radiance-to-reflectance conversion
- Deployment options
- Standalone binary, or service integration with Unibap SCOS
About
Unibap LOOM is a real-time hyperspectral data processing system for satellite operations, running on Unibap’s onboard computers (or standalone) to preprocess raw sensor data directly in orbit. LOOM performs image parsing, band merging, band co-registration, calibration and de-noising, radiance-to-reflectance correction, and orthorectification, producing corrected GEO-TIFF output. It is sensor-agnostic, standardizing image quality across different hardware configurations so that AI models trained on one spacecraft’s data can perform identically on another without retraining. LOOM accepts data directly from sensors or from disk storage in packetized RAW format, supports up to 32 bands (expandable) at up to 4096 pixels width (configurable), uses multi-threaded single-binary processing, and is distributed as .deb packages for Linux. It can run as a standalone binary processing raw sensor data or as a service integrated with Unibap SCOS.
Documentation
Need the full ICD, test reports or a specific revision? Ask the supplier directly.
