NASA's Satellite Image Algorithm Unveils Ancient Cave Art
Original: Technique for Manipulating Satellite Photos Now Reveals Ancient Images (2025)
Why This Matters
Demonstrates how NASA remote-sensing algorithms drive cross-disciplinary breakthroughs in cultural heritage research.
A NASA image-processing technique called decorrelation stretch, originally developed for satellite imagery at JPL, is now widely used to reveal faded ancient rock art and temple paintings. Mathematician Jon Harman adapted it into the DStretch plug-in around 2005, enabling discoveries such as ~200 hidden paintings at Angkor Wat.
NASA's Jet Propulsion Laboratory (JPL) developed decorrelation stretch to extract detail from satellite imagery. The technique remaps colors to an expanded range—using algebraic processes like matrix diagonalization—to heighten subtle contrasts. Around 2005, retired mathematician Jon Harman (PhD, UC Berkeley) discovered a 1996 JPL paper on the algorithm by Ronald Alley, who was then developing requirements for the ASTER instrument launched on the Terra satellite in 1999. Harman, who worked in medical imaging, recognized the algorithm's potential for studying faded rock art—a personal hobby—and built the DStretch plug-in for NIH's open-source ImageJ program, later releasing Android and iOS apps. The technique is grounded in the Karhunen–Loève Transform, co-invented by Michel Loève, one of Harman's former math professors. A notable result: a Singaporean archaeologist used DStretch between 2010 and 2012 to discover approximately 200 previously invisible paintings inside Cambodia's Angkor Wat, including depictions of horseback riders and musicians on chamber walls. The tool has since spread broadly across archaeology and rock art research globally.