AI-Powered Imaging Technique Reveals Hidden Forces on Individual Sand Grains in Dunes

Revolutionary Approach to Granular Physics

Brazilian researchers have reportedly developed a groundbreaking technique that estimates the force exerted on each individual grain of sand within dunes using only images, according to findings published in Geophysical Research Letters. The method combines numerical simulations with artificial intelligence to transform the study of granular system dynamics, potentially unlocking previously unmeasurable physical processes with applications ranging from civil engineering to space exploration.

Understanding Barchan Dunes

The research focused specifically on “barchan dunes,” crescent-shaped structures whose tips orient in the direction of wind or water flow. Sources indicate these distinctive formations appear in remarkably diverse environments, from pipes and riverbeds to terrestrial deserts and even the surface of Mars.

“These dunes appear in very different environments: inside pipes, at the bottom of rivers and seas, in terrestrial deserts, and even on the surface of other planets, such as Mars,” explained Erick Franklin, professor at the State University of Campinas and study coordinator. “This crescent moon shape is an attractor. So, all you need is a reasonable amount of grains on non-erodible soil and a fluid flowing in one direction for a barchan to form.”

Analysts suggest the scales involved vary dramatically—from laboratory aquatic environments as small as 10 centimeters that complete displacement in under a minute, to Martian dunes reaching one kilometer that take approximately a thousand years to move. Despite these differences, researchers maintain the underlying dynamics remain fundamentally similar, making it possible to predict Martian surface evolution from small laboratory dunes.

Overcoming Measurement Challenges

While simply observing dune shape and movement allows inference of wind direction and average intensity, knowing the resulting force on each sand grain has historically been considered impossible. The report states that a laboratory underwater dune might contain 100,000 grains, each just 0.2 millimeters in diameter, while terrestrial desert dunes contain approximately 1 quadrillion grains and Martian dunes reach 100 quadrillion.

“To measure the force acting on each grain, you’d need to place a tiny accelerometer on each one, which simply doesn’t exist,” Franklin noted regarding the measurement challenge., according to industry news

The solution reportedly involved combining laboratory experiments with underwater dunes—which form and move within minutes—with high-resolution numerical simulations that calculate the dynamics of each grain at every instant. These simulations provide force maps unavailable through direct large-scale measurement., according to market developments

AI Implementation and Training

Researchers paired actual dune surface images with force maps from simulations, creating data pairs of images and corresponding force measurements for each grain. Using this data, they trained a convolutional neural network (CNN)—an artificial intelligence model designed for processing spatially structured data like images.

“Based on this, we trained a convolutional neural network to estimate the resulting forces acting on the actual dune grains,” reported Renato Miotto, a postdoctoral researcher at FEM-UNICAMP and visiting researcher at Syracuse University.

Miotto added that “the network was able to infer the distribution of forces from simple images of dunes and even generalize its predictions to shapes it had never seen before.”

Methodological Rigor and Applications

William Wolf, professor at FEM-UNICAMP and study co-author, emphasized the careful data preparation required for the process. “We used high-fidelity, three-dimensional simulations, which allowed us to obtain high resolution of spatial and temporal scales, giving us a level of detail very close to reality,” Wolf stated. “In this way, details of the dynamics and morphology of the dunes were learned by the CNN, and these are essential parameters for the network to be able to generalize to experimental images.”

According to the researchers, the methodology extends beyond sand applications. Miotto suggested that “any granular system that can be seen in an image—whether ice, salt, or synthetic particles—can be analyzed as long as there’s a simulation capable of accurately reproducing the behavior of the material.”

Potential applications reportedly include:

• River silting and beach erosion studies
• Sand movement analysis in ports
• Industrial runoff management
• Martian wind intensity reconstruction from historical images
• Future dune evolution prediction on Mars

“These processes have enormous economic costs and affect entire communities,” Franklin emphasized. “Tools like this can help predict and mitigate damage. In the case of Mars, it’s possible to infer, from widely available images, the intensity of winds in the past and the evolution of dunes in the future.”

Wolf highlighted the collaborative nature of the research, noting that “we’ve been working together for years, combining our expertise in flow physics, fluid mechanics, and computational analysis. It’s an example of how continuous support for basic research can generate advances with impacts in multiple areas.”

References

• https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2025GL116942
• http://en.wikipedia.org/wiki/Grain
• http://en.wikipedia.org/wiki/Dune
• http://en.wikipedia.org/wiki/Mars
• http://en.wikipedia.org/wiki/Sand
• http://en.wikipedia.org/wiki/Desert

This article aggregates information from publicly available sources. All trademarks and copyrights belong to their respective owners.

Note: Featured image is for illustrative purposes only and does not represent any specific product, service, or entity mentioned in this article.