A compact real-time hyperspectral processing system enables anomaly detection directly onboard small UAVs. By converting high-bandwidth spectral streams into immediate geolocated outputs, it removes delays inherent to offline workflows and expands hyperspectral capability to low-SWaP platforms. This allows operators to make rapid, in-situ decisions for environmental, agricultural, defense, and emergency-response missions.
Background: Hyperspectral imaging generates extremely large, high-bandwidth datasets that demand intensive offline processing, often delaying actionable results for hours or days. These requirements have historically confined hyperspectral systems to large aircraft or ground stations, limiting spatial resolution and preventing real-time use in time-sensitive scenarios. Small UAVs lack the onboard compute capacity to process HSI streams without prohibitive latency when offloading data, leaving a critical capability gap for defense, agriculture, environmental monitoring, and emergency response operations that require immediate, on-scene analysis.
Technology Overview: The system streams raw hyperspectral data from a Corning MicroHSI Shark camera over a 1 Gbps Ethernet link to a low-SWaP embedded computer for real-time processing. Incoming ENVI-formatted data is decoded and analyzed using linear or non-linear spectral matching algorithms, including per-pixel hyperspectral unmixing. Spectral detections are geocorrected with UAV navigation and calibration data and projected onto GPS coordinates for immediate visualization at a ground control station. Implemented in C, C++, and Rust, the software delivers onboard real-time hyperspectral analytics that significantly reduce latency relative to conventional offline workflows.
Advantages: • Enables immediate onboard hyperspectral analysis for rapid decision-making • Provides significantly faster processing than conventional or offloaded HSI workflows • Reduces size, weight, and power requirements for small-UAV deployment • Delivers accurate real-time geographic projection of detected spectral signatures • Improves operational efficiency by eliminating offline-processing delays • Supports low-cost embedded hardware for economical field deployment
Applications: • Environmental monitoring and hazard detection • Precision agriculture crop stress analysis • Defense and security reconnaissance • Critical infrastructure inspection • Geological and resource mapping • Emergency and disaster response
Intellectual Property Summary: • United States – 63/884,326 – Provisional – Filed 10/18/2025 – Status: Filed
Stage of Development: Prototype
Licensing Status: This technology is available for licensing.
Licensing Potential: Strong potential for UAV system integrators, environmental and agricultural analytics firms, defense and security contractors, and emergency-response technology providers seeking real-time hyperspectral capabilities on low-SWaP airborne platforms.
Additional Information: Information available upon request.
Inventors: Jayson Boubin, Kenneth Chiu