20 Years of Hyperspectral Processing: From Weeks of Processing to Real-Time Solutions

Over the last two decades, hyperspectral imaging has undergone a transformative evolution in data processing capabilities. Twenty years ago, processing one day of airborne hyperspectral data could take weeks, primarily due to computational limitations and the complexity of processing stages. These stages include handling raw sensor data, processing Inertial Navigation System (INS) information, georeferencing, and performing atmospheric corrections to achieve accurate reflectance results. Only after these steps could application-specific processing, such as classification and analysis, begin.

Significant advancements in computing power, algorithm development, and data handling techniques have dramatically reduced processing times. Innovations like real-time INS integration, improved georeferencing capabilities, and automated atmospheric correction workflows have enabled near-instantaneous reflectance results—a real-time capability that will revolutionize hyperspectral imaging applications.

This presentation explores the key milestones in hyperspectral processing over the past 20 years, highlighting the breakthroughs that have allowed progress from delayed post-processing to real-time solutions. Understanding these advancements provides valuable insights into how technological innovation has shaped modern hyperspectral imaging and why we are able to achieve such efficiency today.

About the presenter
Trond Løke has a master’s degree in photonics from the Norwegian University of Science and Technology (NTNU). He calibrated and characterized Norsk Elektro Optikk's (NEO) first hyperspectral system in 2002 and was part of designing HySpex’s first hyperspectral imaging system in 2003, eventually becoming the CTO. In 2020, Løke became the CEO of NEO located in Oslo, Norway.