We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Computer-Vision Measurement of Turbidity in Large-Volume Liquid Samples: Background Selection, Imaging Geometry, and Preliminary Evaluation Against NTU
AI summary Read the abstract
Researchers tested whether a regular smartphone camera can measure water cloudiness (turbidity), a sign of contaminants, as accurately as pricey lab devices. Early results are promising: photo background color mattered a lot, showing this cheap approach could someday help people check their water quality at home. More testing is needed before it's reliable enough to trust.
Turbidity is conventionally measured using dedicated optical instrumentation that interrogates a controlled sample volume. Camera-based measurement provides a complementary approach in which a substantially larger visible liquid field can be evaluated using intensity, color, spatial, textural, and potentially temporal information. This preliminary study evaluated smartphone-based computer-vision measurement of controlled turbidity conditions using a homogeneous coffee-creamer colloid as an optical surrogate. Six conditions spanning double-filtered water, tap water, and increasing creamer additions were measured using a handheld turbidimeter, producing reference values of 2.7, 8.0, 16.3, 21.0, 29.7, and 43.2 NTU. The same samples were photographed against white and blue backgrounds to investigate whether background selection influences computational measurement performance. Selected optical features were defined before analysis and included global grayscale and RGB intensity, normalized color features, background transmission/contrast, edge energy, and exploratory high-pass energy. Against a white background, mean grayscale intensity decreased with NTU (R² = 0.820; Spearman ρ = −0.943). Against a blue background, edge energy (R² = 0.720; ρ = −0.886) and high-pass energy (R² = 0.720; ρ = −0.771) tracked NTU more strongly than mean intensity (R² = 0.345). Importantly, the blue background did not uniformly improve all measurements. Instead, background selection changed which classes of optical information tracked turbidity most effectively: global intensity measurements performed better against white, whereas edge and high-frequency features performed substantially better against blue. These single-run observations do not establish analytical calibration, accuracy, detection limits, or equivalence to conventional turbidity measurement. They provide preliminary evidence that low-cost camera-derived optical measurements can track conventionally measured turbidity and that background and imaging geometry should be treated as measurement-design variables in computer-vision-based water analysis. Related Papers: 1. Chu MB. Introducing the Spatiotextural Computer Vision Approach to Optical Assays: A Framework for Quantitative Analysis of Assay-Generated Spatial Patterns and Texture. https://doi.org/10.5281/zenodo.22721277 2. Chu MB. Beyond Object Detection: AI for Optical Assays and Real-World Diagnostic Intelligence. Zenodo. 2026. https://doi.org/10.5281/zenodo.19873271 3. Chu MB. Signal Generation and Pattern-Based Detection: Microplastics-Nanoplastics, PCR, and ELISA. Zenodo. 2026. https://doi.org/10.5281/zenodo.19521084 4. Chu MB. Assay Design for Computer Vision: From Chemical Reactions to Spatiotextural Measurement in Water and Other Liquids. https://doi.org/10.5281/zenodo.22729884 5. Chu MB. Computer-Vision Measurement of Suspended Solids: A Preliminary Kaolin Model. https://doi.org/10.5281/zenodo.22732054 6. Chu MB. Spatiotemporal Computer-Vision Measurement of Flocculation and Settling: A Preliminary Kaolin–Polymer Model. https://doi.org/10.5281/zenodo.22733016.
More Papers Like This
Computer-Vision Measurement of Turbidity in Large-Volume Liquid Samples: Background Selection, Imaging Geometry, and Preliminary Evaluation Against NTU
AI summary Read the abstract
Researchers tested whether an ordinary smartphone camera could measure water cloudiness (turbidity), a key sign of contaminants, as accurately as professional lab equipment. Early results show cameras can track cloudiness levels, but accuracy depends on background color and setup. This early-stage work could one day lead to cheaper, at-home water safety testing.
Computer-Vision Measurement of Suspended Solids: A Preliminary Kaolin Model
AI summary Read the abstract
Researchers tested whether ordinary smartphone photos can detect cloudiness in water caused by suspended particles, using clay as a stand-in. Different particle levels produced visibly distinct photo patterns, suggesting smartphone cameras could someday offer a cheap, quick way to screen drinking water quality without lab testing. This is early-stage feasibility work, not a ready-to-use tool yet.
Computer-Vision Measurement of Suspended Solids: A Preliminary Kaolin Model
AI summary Read the abstract
Researchers tested whether an ordinary smartphone camera could detect cloudiness in water caused by added clay particles, as a possible cheap alternative to lab-based water quality testing. Photos clearly showed more cloudiness as clay levels increased, suggesting camera based screening could one day help people monitor water safety quickly and affordably. This is early, preliminary work, not yet a validated testing method.
Assay Design for Computer Vision: From Chemical Reactions to Spatiotextural Measurement in Water and Other Liquids
AI summary Read the abstract
Researchers tested whether smartphone cameras and computer software could detect contaminants like copper or cloudiness in water by analyzing color and texture changes. Early results are promising but still preliminary, this is a proof-of-concept step toward cheaper, faster water testing tools, not a ready-to-use product for checking your tap water yet.
Spatiotemporal Computer-Vision Measurement of Flocculation and Settling: A Preliminary Kaolin–Polymer Model
AI summary Read the abstract
Researchers used simple smartphone photos, taken from the side, to track how water treatment chemicals clear out cloudy particles over time. This early feasibility test suggests cheap camera setups could someday help monitor water treatment quality, making clean water testing more accessible, though the method still needs further validation before real-world use.
Research digests by email
When a large batch of papers lands in the Atlas, we read through it and send a short write-up of what stood out.