The research team, led by Shubham Sharma with Nivedita Kaul and Sumit Khandelwal as co-authors, monitored a working sandstone-carving unit in Rajasthan across three seasons. Rather than relying on a single measurement technique, they combined two complementary approaches. Workers wore personal sampling equipment that captured the respirable fraction of dust, the particles small enough to penetrate deep into the lungs, over their shifts. At the same time, real-time instruments logged concentrations of particulate matter in four size classes: PM10, PM4, PM2.5 and PM1, corresponding to particles with aerodynamic diameters of ten, four, two-and-a-half and one micrometer or less. A 31-channel aerodynamic particle sizer resolved the full size distribution of the airborne dust, while concurrent meteorological measurements recorded the temperature, humidity and wind conditions surrounding each sampling campaign.

The seasonal contrast in fine-particle concentrations was striking. Winter recorded the highest particulate levels of any season, with mean PM2.5 concentrations reaching 109.3 micrograms per cubic meter at the workplace. To put that figure in context, it is roughly an order of magnitude above the annual guideline value recommended by the World Health Organization, and it represents the air in the immediate breathing zone of the artisans rather than a distant ambient monitor. The researchers attribute the winter spike to a combination of factors that converge during the cold months. Temperature inversions and stagnant air suppress the dispersion of dust away from the work area, lower humidity and cooler temperatures alter how particles remain suspended, and the enclosed or semi-enclosed nature of many carving workshops traps emissions close to the source.

Particle size matters as much as particle quantity, and here the study revealed a seasonal fingerprint in the dust itself. During winter, the size distributions showed enhanced accumulation of submicron particles, those smaller than one micrometer, which are the fraction most capable of reaching the deepest regions of the lung and even crossing into the bloodstream. In spring and summer, by contrast, the coarse mode of the distribution grew stronger, reflecting larger fragments that settle more quickly but can still irritate the upper airways. The ratios between size fractions told a consistent story: the contribution of PM1 relative to PM2.5 remained relatively stable across seasons, indicating that once particles are in the fine range, their internal composition shifts little, while the ratio of PM2.5 to PM4 varied more, marking the boundary where seasonal effects reshape the dust cloud.

One of the most technically interesting aspects of the work lies in how the team handled the statistics. The four PM fractions are nested within one another, meaning PM10 includes PM4, which includes PM2.5, which includes PM1. This nesting produces severe multicollinearity: correlations among the fractions exceeded 0.90, making it statistically treacherous to attribute effects to any single size class using ordinary regression. The researchers therefore applied principal component regression, a technique that first compresses the correlated particle-size and meteorological variables into a small set of uncorrelated components and then regresses the outcome on those components. The resulting models explained between 90.4 and 97.7 percent of the variation in PM1 concentrations, an unusually high degree of explanatory power for field exposure data.

The principal component analysis also showed that the drivers of fine-particle concentrations change with the calendar. Associations between PM1 levels and the particle-size and meteorological components varied from season to season, meaning that no single control strategy calibrated in one season can be assumed to work year-round. Strong correlations among all PM fractions pointed to a common source, the mechanical working of the stone itself, but meteorology determines how much of that source ends up in the breathing zone. In winter, the same grinding and chiseling that produces a manageable dust cloud in a breezy spring workshop instead accumulates into a dense, fine-particle haze that lingers around the artisan’s face for hours.

The health stakes of these measurements are not abstract. Sandstone from Rajasthan contains substantial crystalline silica, and inhaling respirable silica dust causes silicosis, a progressive disease for which there is no cure once fibrosis sets in. Studies cited by the authors document high prevalences of silicosis among stone carvers in Brazil, Thailand and Canada’s Nunavut territory, as well as among sandstone mine workers across Rajasthan itself. Indian surveys have reported respiratory symptoms, reduced spirometric readings and radiological abnormalities among stone-cutting workers, and Rajasthan’s own silicosis compensation program has disbursed grants for diagnosed cases and deaths. Previous work by the same research group examined respiratory deposition of particles in stone carving using real-time mass and number concentrations, and the new study extends that line of inquiry by adding the seasonal dimension and the full size-resolved picture.

What makes the findings actionable is their specificity. Because winter emerges as the season of peak exposure, dust-control interventions can be timed and intensified when they matter most. The literature on stone fabrication points to several engineering controls that have proven effective elsewhere: on-tool shrouds and local exhaust ventilation that capture dust at the point of generation, wet methods that suppress dust before it becomes airborne, and enclosure of grinding stations. The study’s seasonal size distributions offer a further clue for control design, since a system tuned to capture coarse particles in summer may underperform against the submicron accumulation mode that dominates in winter. Administrative measures, such as rotating workers away from grinding tasks during high-exposure periods and ensuring proper use of respiratory protection, can also be scheduled around the winter peak.

The authors are careful to frame the scope of their conclusions. The measurements come from a single sandstone-carving workplace monitored over three seasons, and they note that larger multi-site and longer-duration investigations are required to establish how broadly the seasonal patterns apply across the stone-carving sector. Rajasthan’s carving industry is vast and largely informal, ranging from heritage-restoration workshops to small units producing export-quality decorative stonework, and working conditions can differ substantially between sites. Nonetheless, the study provides something the sector has lacked: a seasonally resolved, size-resolved characterization of what carvers actually breathe, grounded in personal sampling rather than area monitors alone.

For the artisans of Rajasthan, the message embedded in the data is quietly urgent. The dust that glitters in a shaft of winter sunlight is not merely a nuisance but a precisely measurable hazard whose finest particles concentrate exactly when the weather conspires to keep them airborne and close. The study’s models, explaining more than ninety percent of the variance in fine-particle concentrations, demonstrate that this hazard is predictable, and what is predictable can be managed. As India’s natural stone industry faces growing scrutiny over labor conditions and occupational disease, research of this kind supplies the evidence base on which targeted regulation, engineering investment and worker protection can finally be built, season by season and micron by micron.

Subject of Research: Seasonal particulate matter exposure and particle size distribution among sandstone-carving workers in Rajasthan, India

Article Title: Seasonal variation in particulate emissions, particle size distribution, and worker exposure at a sandstone-carving workplace in Rajasthan

Article References: Sharma, S., Kaul, N., & Khandelwal, S. (2026). Seasonal variation in particulate emissions, particle size distribution, and worker exposure at a sandstone-carving workplace in Rajasthan. Air Quality, Atmosphere & Health, 19(10), Article 217. https://doi.org/10.1007/s11869-026-02110-5

Image Credits: AI Generated

DOI: 10.1007/s11869-026-02110-5

Keywords: particulate matter, respirable dust, silica exposure, silicosis, sandstone carving, occupational health, Rajasthan, PM2.5, particle size distribution, principal component regression, air quality, worker exposure

The core of the method is a two-stage pipeline. First, photographs of rockfill material—whether spread on a laboratory bench or compacted in layers on a dam site—are fed into Cellpose, a generalist deep learning algorithm originally developed for segmenting individual cells in microscopy images. Cellpose’s neural network learns to delineate the boundaries of touching, overlapping, irregularly shaped objects, which makes it unusually well suited to the visual chaos of a pile of quarried rock. The algorithm produces a high-precision segmentation mask in which every individual particle is separated from its neighbors, solving the classic bottleneck that has limited automated rock image analysis: stones that touch or partially overlap are notoriously difficult for conventional thresholding techniques to separate.

Once Cellpose has produced the segmented image, the researchers turn to the MorphoLibJ plugin of ImageJ, a mature mathematical morphology library, for post-processing and quantitative computation. From each segmented particle, the software extracts geometric features including area, particle size, and aspect ratio. Aggregated over the hundreds or thousands of stones visible in a single image, these features yield the particle size distribution curve—one of the most fundamental descriptors of any granular construction material—without the sieving, weighing, and labor that traditional gradation testing demands. The team reports that the ImageJ-Cellpose method could accurately determine both the particle size distribution curve and the porosity of rockfill materials, results that were validated against conventional screening and porosity experiments conducted in the laboratory.

Beyond reproducing established metrics, the study introduces new shape parameters for rockfill material, including what the authors call the long-short ratio and the area ratio. Shape matters in geotechnical engineering more than casual observers might expect: angular, elongated, and flat particles interlock differently than rounded, equidimensional ones, influencing shear strength, particle breakage under load, and the volumetric behavior of the compacted mass. By statistically analyzing the distribution of these new shape descriptors across particle populations, the researchers offer a quantitative framework for describing rockfill morphology that goes beyond simple size gradation, opening a path toward correlating shape statistics directly with mechanical performance.

The critical question for any laboratory method is whether it survives contact with the real world. To answer it, the team partnered with a pumped-storage power station, where layered rolling experiments were carried out during actual rockfill placement. As the material was compacted by rollers in successive passes, images were collected at each stage and processed through the intelligent recognition pipeline. The results showed that the method could effectively trace how the particle size distribution and porosity of the rockfill evolved with each rolling pass—precisely the kind of real-time feedback that compaction quality control has historically lacked. Encouragingly, the trends identified by the image-based approach aligned well with those derived from the pit measuring method, the traditional but destructive and labor-intensive standard for evaluating compaction in the field.

The significance of this validation lies in what it could change about dam construction practice. Rockfill dams are among the largest man-made structures on Earth, and their long-term safety depends on how well the fill is compacted during construction; poor compaction leads to post-construction settlement, cracking of upstream concrete face slabs, and heightened vulnerability to overtopping failure and seismic loading. Current quality control typically relies on periodic pit sampling, in which crews excavate a hole, weigh the removed material, and measure density—a slow process that samples only a tiny fraction of the dam’s volume. A camera, a laptop running Cellpose and ImageJ, and the proposed analytical workflow could in principle assess compaction indicators continuously and across the entire working surface, converting construction monitoring from intermittent spot checks into comprehensive surveillance.

Technically, the study is notable for transplanting tools from an unexpected domain. Cellpose was created by biologists to solve the problem of segmenting cells in diverse image types without retraining a network for each new dataset, and its generalist design proved transferable to geological imagery. ImageJ, likewise, grew out of biomedical imaging but ships with general-purpose morphology operators that apply equally well to mineral particles. By connecting a state-of-the-art segmentation network to a battle-tested analysis library, the Xi’an team sidestepped the need to build bespoke deep learning infrastructure or large annotated rockfill datasets from scratch, lowering the barrier to adoption for engineering firms and construction supervisors who cannot maintain their own machine learning teams.

The method also complements a growing body of work on computer vision in geotechnical monitoring. Prior studies have used video image recognition to detect rockfill gradation, deep learning segmentation to calculate gradation from photographs, and instance segmentation to predict gradation for engineering projects. What distinguishes the new approach is its end-to-end coverage: rather than estimating only particle sizes, it simultaneously delivers size distribution, individual particle shape metrics through newly proposed parameters, and porosity, all from ordinary images, and it does so in a form validated against both sieving tests in the laboratory and pit measurements on an active construction site. The authors argue that this makes the technique suitable for real-world engineering applications and a robust means of ascertaining particle size information during the construction of rockfill dams.

Looking forward, the research points toward a swift and precise evaluation method for the compaction quality of rockfill material—a capability the authors describe as paramount for the comprehensive understanding and analysis of the safety conditions of rockfill dams and similar projects. If deployed at scale, the ImageJ-Cellpose workflow could give dam engineers a continuous, quantitative picture of how each roller pass is reshaping the granular skeleton of an embankment, catch gradation or compaction anomalies before they are buried under the next lift of rock, and feed high-quality geometric data into the numerical models used to predict settlement, seepage, and seismic response. In an era when aging dam infrastructure faces intensifying climate and seismic stresses, the idea that a deep learning algorithm designed for counting cells can help safeguard structures holding back billions of liters of water is a striking demonstration of how tools migrate across scientific boundaries—and of how much of modern construction safety may soon rest on image recognition pipelines that work quietly, frame by frame, on the rocks beneath the rollers.

Subject of Research: Intelligent image-based geometric characterization of rockfill materials using deep learning segmentation for dam compaction analysis

Article Title: Intelligent geometric characterization of rockfill materials using ImageJ-Cellpose: a novel approach for accurate morphological analysis

Article References: Ma, C., Zhou, C., Hou, Y., & Cheng, L. (2026). Intelligent geometric characterization of rockfill materials using ImageJ-Cellpose: a novel approach for accurate morphological analysis. Cluster Computing, 29(13), Article 779. https://doi.org/10.1007/s10586-026-06480-4

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06480-4

Keywords: rockfill materials, deep learning, Cellpose, ImageJ, image segmentation, particle size distribution, porosity, particle shape, compaction quality, rockfill dams, image recognition, geotechnical engineering

The research centers on a new material called Colorado School of Mines Lunar Highlands Type-Testbed simulant, or CSM-LHT-T, a highland-type regolith analog produced by the university’s Space Resources Program. Roughly 110 metric tons of a final blend, designated CSM-LHT-T-30, now fill the newly constructed Mines Lunar Surface Simulator testbed, a large-scale facility designed to host mid-Technology Readiness Level experiments in conditions more realistic than laboratory bench tests can offer. Two prototype mixtures, CSM-LHT-T-25 and CSM-LHT-T-30, were evaluated during the study, and the denser, better-graded 30 blend was selected as the testbed infill after its measured properties proved closer to Apollo-era estimates of actual lunar highland soil.

The team, led by Ian E. Jehn along with colleagues at the University of Oklahoma, Slate Geotechnical Consultants, and the architecture firm Skidmore, Owings & Merrill, subjected the simulants to a battery of standardized American Society for Testing and Materials procedures. These included sieve analysis and laser diffraction for particle size distribution, Scott volumeter and Proctor compaction tests for minimum and maximum dry density, one-dimensional consolidation testing over a fourteen-day regime for the compression index, and both triaxial and direct shear tests to determine cohesion and friction angle. Reference mare simulants BP-1 and JSC-1A were run through the same procedures to verify that the testing methods themselves were sound, since the ultimate benchmark was the geotechnical record assembled from Luna, Lunokhod, Surveyor, and Apollo mission data, supplemented by more recent orbiters.

The study’s central methodological argument is that particle size distribution, long the headline metric in simulant marketing and comparison tables, is fundamentally insufficient for judging whether a simulant will replicate the mechanical behavior of real regolith. Particle geometry, surface roughness, mineralogy, soil fabric, and above all density all exert powerful influences on shear strength, compressibility, and deformation response. Laboratory work on crushed terrestrial sands has shown that soils with essentially identical gradations can exhibit markedly different shear strengths, and the authors note that recent research has demonstrated simulants with matching particle size distributions diverging in geotechnical behavior purely because of differences in density state.

Density, in fact, emerges as the study’s recurring theme. Friction angle and cohesion are not fixed constants of a granular material; both rise as the material is compacted. On the Moon, regolith density increases with depth as overlying material presses particles into tighter interlock, a relationship documented in the classic cohesion and friction angle curves derived from Apollo and Lunokhod measurements. The authors warn that any reported simulant property that omits the density at which it was measured is of limited engineering value, and may even hint at uncontrolled sample preparation. For a testbed that must simulate the stress-dependent behavior of a lunar surface under lander legs, excavation blades, or rover wheels, mismatched density profiles translate directly into mismatched predictions of bearing capacity and settlement.

When the team compared CSM-LHT-T-30 against estimates for actual lunar highland regolith, the results were encouraging on most fronts. The simulant’s particle size distribution falls largely within one standard deviation of the lunar average established by Carrier, and it carries the same classification, well-graded under the Unified Soil Classification System and Highland Medium under a newer lunar-specific classification scheme, as Apollo highland core samples. Differences in density, compression index, and friction angle relative to lunar highland regolith estimates all came in below roughly 30 percent. In a settlement model applying a 10,000-newton load over a one-meter footprint to a simulated 200-centimeter regolith column, CSM-LHT-T-30’s predicted surface deflection differed from the lunar estimate by just over 9 percent, far outperforming BP-1 at about 84 percent and JSC-1A at about 38 percent.

One number, however, stands out. The simulant’s cohesion is substantially higher than estimated values for real highland regolith, a discrepancy the authors estimate could translate into cohesion forces on the actual lunar surface being lower by as much as a factor of several. Notably, CSM-LHT-T-30 is not an outlier; its cohesion sits within a similar range to other highland simulants currently in circulation, exceeding the widely used OB-1A by only about 15 percent. This suggests the discrepancy may be systemic across the simulant industry, potentially rooted in differences in mineralogy or particle morphology at the microscale. The team also cautions that their shear tests used relatively low normal loads, appropriate for construction-scale analysis, and that further testing across a broader stress range will be needed to pin down the failure envelope more precisely.

The authors are careful to scope their claims. CSM-LHT-T-30 replicates only the immediate, relatively homogeneous top layer of lunar regolith; it does not incorporate the rocks and stratigraphic complexity found at depth, and the measured properties should be used for simulant comparison and relative testbed performance rather than for structural design or in situ excavation modeling. They also flag a practical concern that rarely appears in simulant literature: the industry typically does not retire or re-characterize material after use. Simulants in testbeds are reused for years, during which operational traffic can segregate particle sizes, fracture grains, and alter packing, quietly shifting the very properties the facility was calibrated around. With more than 100,000 kilograms of simulant in the new testbed, spatial and temporal variability is a live research question.

The path forward outlined in the paper includes characterizing the simulant at multiple density states, since all shear and consolidation values in this study were anchored to the 90 percent relative density specified as a terrestrial construction minimum under the International Building Code, a datum useful for comparison but not necessarily representative of the Moon’s looser uppermost surface. The team also plans systematic sampling of the testbed at various locations and depths to map heterogeneity and inform maintenance decisions such as periodic remixing, and microscale investigations into the cohesion gap. For a field racing to de-risk lunar construction before hardware ever leaves Earth, the message is clear: simulant fidelity is a multi-parameter problem, and density-specific reporting should become the community’s non-negotiable standard.

Subject of Research: Geotechnical characterization of lunar regolith simulants for testbed experimentation and lunar surface engineering analysis

Article Title: Implications of lunar simulant geotechnical properties on testbed experimentation and engineering analysis and reported properties of Colorado School of Mines highland simulant

Article References: Jehn, I. E., Casasbuenas Cabezas, Y. I., Bounds, T. D., Houston, G. G. N., Dreyer, C. B., Johnson, C., Murphy, D., Smith, S., Williams, T., Caluk, N., & Lee, P. (2025). Implications of lunar simulant geotechnical properties on testbed experimentation and engineering analysis and reported properties of Colorado School of Mines highland simulant. Space and Planetary Resources, 1(1), Article 5. https://doi.org/10.1007/s44461-025-00002-7

Image Credits: AI Generated

DOI: 10.1007/s44461-025-00002-7

Keywords: lunar regolith simulant, geotechnical properties, Colorado School of Mines, CSM-LHT-T, testbed experimentation, shear strength, particle size distribution, soil density, Artemis program, lunar construction, simulant fidelity, cohesion and friction angle