Smartphone Sun Photometry for Aerosol Depth

ISEF Category: Earth and Environmental Sciences

Subcategory: Atmospheric Science  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: Full Year

The Hook

Haze changes how much sunlight reaches your eyes, but your phone can still help measure it. A tiny pinhole and a color analysis app can turn sunset photos into data. That means you can study air quality with tools already in your pocket.

What Is It?

Aerosol optical depth, or AOD, is a number that describes how much particles in the air block or scatter sunlight. Think of it like looking through clean glass versus frosted glass. The frosted version hides more of what is behind it, and aerosols do something similar to the sun’s light.

Sun photometers measure how bright the sun looks in different wavelengths. A cheap smartphone setup can mimic part of that idea by using a pinhole to control glare and RGB analysis to measure color changes in sunset or near-sunset images. RGB means red, green, and blue channel values from a photo. Those channel changes can track how much light the atmosphere has filtered.

The calibration step matters. AERONET, a global network of standardized sun photometers, gives reference AOD data at many stations. You can compare your phone-based measurements against nearby AERONET records, then test whether machine learning can correct the errors caused by haze type, cloud edges, or changing lighting conditions.

Why This Is a Good Topic

This project works well for science fair research because you can test a real atmospheric measurement problem with accessible tools. You are not just taking pictures, you are trying to convert photos into a number that connects to air quality and visibility. That gives you a clear variable, a real-world use, and plenty of room for analysis. You can also compare different haze conditions, calibration methods, or models, which makes the project more original.

Research Questions

• How does smartphone RGB analysis compare with AERONET AOD values at nearby times?
• What is the effect of haze type on the error between phone-based AOD estimates and reference data?
• Does a pinhole attachment improve the repeatability of sunset photometry measurements?
• To what extent do different color channels predict aerosol optical depth under smoky, dusty, or humid conditions?
• Which machine learning model best corrects phone-based AOD estimates using local weather and image features?
• How does sun angle at capture time affect the accuracy of smartphone sun-photometry readings?
• What is the effect of image exposure settings on the stability of AOD estimates from RGB data?

Basic Materials

• Smartphone with manual camera controls or a camera app that saves exposure data.
• Small pinhole attachment or cardboard pinhole mount.
• Dark material or 3D-printed holder to block stray light.
• Tripod or stable phone stand.
• Notebook or spreadsheet for field notes.
• Access to AERONET station data online.
• Free image analysis software such as ImageJ.
• Internet-connected computer for data analysis.

Advanced Materials

• Smartphone with raw image capture support.
• Custom pinhole or sunglass-style optical mount for repeatable alignment.
• Portable tripod with angle markings.
• Light meter or sky radiometer if available for cross-checking.
• GPS-enabled field log or timestamped weather station data.
• Access to AERONET station records and local meteorological data.
• Python or R for calibration and model fitting.
• Statistical software for regression, cross-validation, and error analysis.

Software & Tools

• ImageJ: Measures RGB intensity from your photos and helps standardize image analysis.
• Python: Lets you clean data, fit calibration curves, and test machine learning correction models.
• R: Helps you run regression models, uncertainty checks, and visual comparisons.
• Google Sheets: Tracks field observations, timestamps, and reference values in one place.
• QGIS: Maps sampling locations and compares your readings with air quality or station data.

Experiment Steps

1. Define the exact atmosphere variable you want to estimate, then choose the narrowest possible field conditions you can compare across days or sites.
2. Design a repeatable phone-and-pinhole geometry so each photo uses the same viewing angle, framing, and light-blocking setup.
3. Match your phone measurements to a reference source such as AERONET, then decide how close in time and location each pair must be.
4. Build a calibration curve that converts color or brightness features into an aerosol estimate, then test whether the relationship stays stable.
5. Separate your data by haze type, sun angle, or weather condition, then plan a correction model for the cases where the simple calibration fails.
6. Compare model performance with a clear error metric, then decide which version generalizes best to new days or new skies.

Common Pitfalls

• Shooting photos with different exposure settings, which makes RGB values change for camera reasons instead of aerosol reasons.
• Comparing phone data to AERONET values taken too far apart in time, which can hide real atmospheric changes.
• Using sunset images with thin clouds near the sun, which can distort the light path and bias the estimate.
• Ignoring the phone’s automatic white balance, which can shift color channels and break your calibration.
• Mixing haze types in one calibration model, which can make the algorithm look accurate in training data but fail on new skies.

What Makes This Competitive

A strong project here does more than match a reference station. You need a careful validation plan, a clear uncertainty estimate, and a reason your method works across different haze conditions. A more competitive version compares multiple correction methods, tests several feature sets, and reports where the phone model breaks down. If you can show when the method works, when it fails, and why, your project looks much closer to real atmospheric research.

Project Variations

• Test whether a phone sun photometer works better at sunrise than sunset because of changing sky brightness.
• Compare RGB-based aerosol estimates with a grayscale brightness method to see which predicts AOD more accurately.
• Add weather data such as humidity, wind, or smoke reports to see whether a correction model improves local haze estimates.

Learn More

• AERONET: Search the NASA AERONET site for station data, methods, and calibration background.
• NASA Earthdata: Search for aerosol, atmosphere, and remote sensing articles and datasets.
• NOAA Air Resources Laboratory: Find background on aerosol transport, visibility, and atmospheric optics on NOAA pages.
• PubMed: Search for review articles on smartphone air-quality sensing and aerosol optical depth estimation.
• Atmospheric Measurement Techniques: Search the journal for papers on sun photometry, calibration, and aerosol retrieval methods.
• MIT OpenCourseWare: Search for atmospheric science and remote sensing lecture materials that explain radiation and scattering.