By This Hour Science Desk
A digital-image technique first used in NASA-related satellite work is being applied to a very different record of the past: paintings and markings so faded that visitors can pass them without recognizing that they are there.
NASA says the method, called decorrelation stretch, has helped make low-contrast features in photographs more visible and has found an audience among archaeologists studying ancient rock art and historical imagery. Its most striking cited example lies in Cambodia’s Angkor Wat, where faint paintings in the temple’s central tower include horseback riders and a traditional musical ensemble. The reported findings matter because the tool does not physically alter a site; it changes how existing digital image data are displayed, potentially directing attention to details that ordinary viewing misses.
The account illustrates a longer path from Earth observation to archaeological interpretation. A method intended to extract more information from satellite scenes was adapted in software for close-range images of walls, caves, human remains and landscapes. That crossover broadens the range of material researchers may be able to inspect, but it also leaves a crucial distinction intact: enhancing an image can reveal a feature worth investigating; it does not by itself settle what the feature is, when it was made or what it means.
Contrast, rather than reconstruction, is the central function
Decorrelation stretch works by heightening contrasts in digital imagery, making visual differences more apparent. In practical terms, a faint painted line or patch of pigment that blends into stone, plaster or another background can become easier to distinguish after processing. The technique is therefore an aid to seeing patterns already represented in an image, rather than a substitute for field examination or archaeological judgment.
That difference is important when dealing with worn cultural material. A clearer-looking digital result may help an investigator identify an area for closer study, compare photographs, map possible imagery or document a fragile surface. Yet a visually compelling result still has to be assessed against the original object and the surrounding evidence. An enhanced image cannot, on its own, establish whether a marking is deliberate, how old it is, or whether an apparent figure has been interpreted correctly.
NASA describes the method as particularly popular in the study of ancient rock art. This use follows naturally from the condition of much such material: pigments and outlines can be difficult to perceive because of fading, surface change, lighting and the limits of ordinary photography. The reported range of uses goes beyond rock art, however. The technique has been associated with examinations of imagery at sites in Norway, Egypt and Canada, as well as with work involving buried remains of ancient Greek buildings and tattoos on mummified human remains.
Those examples point to a common problem rather than a single type of artifact. Archaeologists often confront evidence that is present but visually subdued. A digital tool that separates subtle differences may give them a way to locate and record candidates for further work. The value claimed for decorrelation stretch is thus not that it replaces archaeological methods, but that it may increase the chance that faint visual information enters the investigation at all.
Angkor Wat offers a vivid case of details hidden in plain sight
At Angkor Wat, the NASA account says, an archaeologist identified faded paintings between 2010 and 2012. Among them were depictions of horseback riders and a traditional musical ensemble positioned high in the complex’s central tower. NASA says the work also identified roughly 200 other paintings across the temple complex.
The account emphasizes the unusual visibility problem posed by the paintings: large numbers of visitors could move through the monument without noticing them because they had faded so severely. That makes the case a useful illustration of why a contrast-enhancement technique can be consequential. The immediate gain is not the creation of a new object or inscription, but a different view of visual traces that were already on the surfaces of an intensively visited historic place.
Even so, the reported discovery should not be read as a claim that software alone produced a complete historical interpretation. The supplied account identifies the subjects of particular paintings and gives a broad count for others, but it does not provide further detail here about their dating, preservation condition, authorship or scholarly conclusions. Nor does it describe what independent methods were used to confirm each identification. Those omissions limit what can responsibly be inferred from the example.
For cultural-heritage work, that limitation is as meaningful as the technical advance. A processed image may be an effective first signal, particularly when imagery is nearly invisible to the naked eye. But a signal requires careful follow-up. Context within a building, comparison with the unprocessed photograph, the physical surface itself and specialists’ interpretation all bear on whether an enhanced form should be treated as a painting, an incidental feature or something else.
From a satellite instrument to an accessible plug-in
The reported lineage of the method begins with work at NASA’s Jet Propulsion Laboratory in Southern California. NASA says Ronald Alley described the underlying algorithm in a 1996 paper while developing applications for the Advanced Spaceborne Thermal Emission and Reflection Radiometer, known as ASTER. ASTER was a Japanese imaging instrument aboard NASA’s Terra satellite, and the effort concerned drawing information from satellite imagery.
According to NASA’s account, a former supervisor of Alley’s at JPL had co-invented decorrelation stretch, while Alley recognized possible uses for it with ASTER imagery. The history matters because it frames the method as an analytical response to the information limits of remotely sensed images. Satellite scenes can contain subtle variations that conventional presentation does not make easy to inspect. The same broad challenge appears in photographs of faded cultural surfaces, even though the scale and subject of the images are radically different.
Jon Harman later translated that possibility into a tool for archaeological imagery. NASA says Harman, a rock-art enthusiast with experience in medical imaging, saw examples of Martian images before and after decorrelation stretch around 2005. The difference in visible detail led him to consider the technique’s possible use on faint ancient imagery. He then adapted it into DStretch, a plug-in for ImageJ, an open-source image program developed by the National Institutes of Health.
This sequence is significant because it made the approach usable beyond the setting in which the algorithm had been discussed. The relevant transfer was not simply from space science to archaeology; it also involved putting a computational method into a form that people working with their own photographs could apply. NASA presents Harman’s professional background in medical imaging as part of the explanation for why he could make that adaptation.
During DStretch’s development, Harman applied it to an image from the Cave of San Borjitas in Baja California, Mexico, according to the NASA page. The processed result made a yellow figure visible near the center of the image, an outcome that he took as evidence that the approach had practical value. The example captures the attraction of the method: it may bring a possible image to attention where the original photograph offers few obvious clues.
Wider adoption brings both access and interpretive responsibility
NASA says Harman receives about 200 requests for DStretch each year. It also says he created smartphone applications around 2010 that use a shortcut intended to mimic decorrelation stretch, and that those applications have been downloaded thousands of times. If accurate, those figures suggest that the technique is not confined to a small specialist setting.
Broader access has an evident benefit. Field researchers, heritage workers and others interested in faint imagery may be able to test photographs without needing the original satellite-imaging environment. It could speed the preliminary identification of areas that deserve attention and give people a way to make comparisons across image sets. The reported publication of papers on DStretch’s archaeological usefulness indicates that the approach has been discussed in archaeological contexts, though the supplied material does not identify those papers or summarize their findings.
Accessibility also makes disciplined interpretation more important. A tool that increases visual contrast can produce results that appear more definite than the underlying evidence warrants, particularly to an untrained user. The central question is not only whether a color separation or shape appears after processing, but whether it corresponds to a real, culturally meaningful feature on the material being studied. Responsible use requires retaining originals, recording processing choices and testing apparent discoveries against direct examination and relevant archaeological context. The NASA account provided does not set out a validation protocol, so it cannot establish how consistently such safeguards have been applied in the cited cases.
The source likewise gives no comparative measure of how often the technique produces useful identifications, no account of false positives and no detail on the condition of every site mentioned. It supports a narrower conclusion: NASA describes decorrelation stretch and DStretch as tools that have been used to make faint visual details easier to see across several archaeological applications. It does not support a conclusion that every enhanced pattern represents a verified discovery.
NASA’s description is the sole supplied basis for this report, and the reported uses, counts and historical details have not been independently corroborated here. That uncertainty does not negate the reported technique or its potential; it defines the limits of the available evidence. The most durable significance of the account may be its demonstration that methods developed to read subtle signals in satellite data can be repurposed to guide attention toward vulnerable traces of human history—provided enhanced images remain the beginning of inquiry, not its final proof.
For further context on this subject, see Three hikers rescued on Mount Shasta after reported Gemini trip planning.
Reporting notes
What is confirmed: NASA describes reported uses at Angkor Wat and archaeological sites in several countries.
Why this matters: The method may help locate fragile visual evidence without altering the original site, while requiring careful verification.
What remains unclear: The supplied account does not provide independent confirmation, validation rates or detailed methods for each cited identification. This report is based on one source and has not been independently corroborated.
