A lightweight onboard software system processes hyperspectral data in real time on small UAVs, converting raw spectral streams into actionable, geolocated classifications. By removing the need for slow offline workflows, it enables immediate field decisions for defense, agriculture, environmental monitoring, and emergency response, bringing advanced hyperspectral intelligence to low-SWaP platforms.
Background: Hyperspectral imaging is typically processed offline, limiting its usefulness in missions where rapid decisions are critical. Existing real-time systems are designed for satellites or large aircraft and are not adapted for the size, weight, power, and cost constraints of small UAVs. Current UAV-integrated solutions frequently support only basic preprocessing, still requiring post-flight analysis to obtain meaningful outputs. Integration complexity, high system cost, and lack of onboard real-time detection and geospatial projection restrict broader adoption of hyperspectral sensing across time-sensitive defense, agricultural, environmental, and emergency-response applications.
Technology Overview: The software operates on a low-SWaP microprocessor connected to a pushbroom hyperspectral camera, reading high-rate ENVI-format data streams at up to 300 Hz. It decodes raw spectral measurements into radiance values and applies user-defined detection functions for real-time per-pixel classification. By fusing UAV and camera sensor data, the system performs inverse stereographic projection to geolocate classified pixels on the Earth’s surface. Built in modular C++20, the framework is hardware-agnostic and compatible with any HSI camera capable of runtime data offloading, providing a flexible platform for mission-specific hyperspectral analytics.
Advantages: • Enables immediate hyperspectral detection and geospatial projection onboard UAVs • Supports multiple hyperspectral cameras through a hardware-agnostic architecture • Operates on low-SWaP processors suitable for consumer UAV platforms • Allows integration of custom detection algorithms for mission-specific needs • Provides real-time pixel-level geolocation using integrated UAV sensor data
Applications: • Emergency response and public safety monitoring • Precision agriculture and forestry assessment • Defense and law enforcement surveillance • Environmental and water quality monitoring • Critical infrastructure inspection • Geological and mineral mapping
Stage of Development: Prototype
Licensing Status: This technology is available for licensing.
Licensing Potential: Well-suited for UAV manufacturers, remote sensing firms, agricultural analytics providers, defense contractors, and emergency-response technology developers seeking real-time hyperspectral intelligence on resource-constrained airborne platforms.
Additional Information: Information available upon request.
Inventors: Jayson Boubin