Admin login
UF/IFAS, University of Florida Water Resources Research LabHydrology & Irrigation Research
GeoDiff-SR: 4x diffusion super-resolution for agricultural drone imagery

Live demoGenerative AI · Remote sensing

GeoDiff-SR

4x diffusion super-resolution for agricultural drone imagery

GeoDiff-SR is a conditional denoising-diffusion model that enhances the resolution of agricultural UAV images by a factor of four while keeping reconstructions faithful to the underlying scene. It recovers fine canopy detail without the hallucinated texture typical of GAN-based super-resolvers, so agronomic measurements stay reliable.

The model was trained on 2,136 sweet corn and green bean images collected over three growing seasons (2020 to 2023) at the UF/IFAS Tropical Research and Education Center. Try it on bundled low-resolution plot frames in your browser.

  • Mean opinion score 0.98, vs 0.82 for Real-ESRGAN and 0.45 for bicubic
  • Canopy-coverage error 0.71%, vs 2.37% for Real-ESRGAN
  • Trained on 2,136 sweet corn and green bean UAV images (2020-2023)
  • Rated in a blind human-perception study with 38 participants
PID-LDM: RGB to thermal: Canopy temperature and crop water stress (CWSI) from low-cost RGB drone photos

Live demoGenerative AI · Crop water stress

PID-LDM: RGB to thermal

Canopy temperature and crop water stress (CWSI) from low-cost RGB drone photos

PID-LDM is a physics-informed latent diffusion model that synthesizes canopy thermal imagery directly from RGB UAV photographs, then turns the predicted temperature into a Crop Water Stress Index (CWSI) map. It brings thermal-camera-grade water-stress monitoring within reach of RGB-only drone missions.

The live demo runs a Pix2Pix baseline on bundled plot images; the published PID-LDM model reaches a 0.90 °C mean absolute error.

  • 0.90 °C mean absolute error on 271 held-out images
  • CWSI error 0.011, an order of magnitude below GAN baselines
  • Trained on 2,710 paired RGB and thermal UAV images
  • From RGB capture to irrigation-ready CWSI maps
Soil Moisture Dashboard: Soil sensor and weather analytics that run in your browser

LiveSensors · Weather · Machine learning

Soil Moisture Dashboard

Soil sensor and weather analytics that run in your browser

The Soil Moisture Dashboard turns soil-sensor logger exports and weather data into analysis-ready information. Combine logger files into one table, bring in FAWN weather station data, fuse the sources, model evapotranspiration and forecast soil moisture, with machine learning that runs in the browser.

  • Combine soil-sensor logger files into one table
  • Live FAWN weather station data
  • Data fusion, evapotranspiration modeling and forecasting
  • Machine learning that runs in the browser
Sabal Field Advisor: Irrigation decisions for Florida fields, with a receipt for every recommendation

Testing (access code)Web app · Irrigation decision support

Sabal Field Advisor

Irrigation decisions for Florida fields, with a receipt for every recommendation

Sabal Field Advisor brings public data for a Florida field together in one place: FAWN and National Weather Service weather and forecasts, USDA soil surveys, USGS groundwater wells, drought and flood maps, and state water rules. A daily FAO-56 root-zone water balance says whether to irrigate or skip and when the next irrigation is due, and every recommendation comes with a receipt that lists its sources, data dates and steps. Growers can draw their fields, add their own records and ask questions in plain language. The web app is a staging site for testing and asks for an access code. Android and iPhone apps are also in testing.

  • Irrigate-or-skip call from a daily FAO-56 root-zone water balance
  • Weather, forecasts, soils, groundwater and water rules for each field
  • A receipt for every recommendation: sources, data dates and steps
Shamba: Rainy-season advice for farmers in Ethiopia, Kenya and Tanzania

Testing (access code)Web app · Rainfall and seasons

Shamba

Rainy-season advice for farmers in Ethiopia, Kenya and Tanzania

Shamba turns public satellite rainfall and climate data into season information for any point in Ethiopia, Kenya and Tanzania: when the rains usually start, whether they have started, false starts, dry spells, and which crops the season can carry. Pick a place on the map or search for a town or village. Short answers come in English, Amharic, Afaan Oromoo or Kiswahili, and every number shows its source and data date. The web app is a staging site for testing and asks for an access code. Android and iPhone apps are also in testing.

  • Season start, false starts and dry spells from CHIRPS and TAMSAT satellite rainfall
  • Crop water outlook for maize, sorghum, wheat and other crops
  • Answers in English, Amharic, Afaan Oromoo and Kiswahili
NISAR x FAWN: NISAR satellite soil moisture checked against FAWN stations in Florida

Dashboard · Satellite soil moisture

NISAR x FAWN

NISAR satellite soil moisture checked against FAWN stations in Florida

NISAR x FAWN is a dashboard that validates soil moisture from the NASA-ISRO NISAR satellite (the SME2 product) against Florida Automated Weather Network (FAWN) ground stations. A Florida map shows which stations each satellite pass covers, and the dashboard pairs every pass with the station readings and reports how closely they agree. FAWN's live 15-minute data and NASA's pass search sit beside the validation release, and passes published after the release are processed with the same code and marked as not yet validated. Sign in to use it; ask the lab for an account.

  • NISAR SME2 soil moisture compared with FAWN station readings
  • Florida coverage map of the stations each satellite pass covers
  • Live FAWN data and new NISAR passes, clearly marked as not yet validated
OpenET x Florida: OpenET evapotranspiration checked against eddy covariance towers and FAWN stations in Florida

Dashboard · Evapotranspiration

OpenET x Florida

OpenET evapotranspiration checked against eddy covariance towers and FAWN stations in Florida

OpenET x Florida is a dashboard that compares daily evapotranspiration from OpenET with measurements at five AmeriFlux eddy covariance towers in Florida, and OpenET reference evapotranspiration with 49 Florida Automated Weather Network (FAWN) stations. It shows the results for each OpenET model and the ensemble, by month and by year, and the daily data can be downloaded as CSV files. Sign in to use it; ask the lab for an account.

  • Daily OpenET evapotranspiration compared with eddy covariance tower measurements
  • OpenET reference evapotranspiration compared with FAWN station data
  • Results by model, month and year, with daily data downloads

Using our tools in your work?

We welcome feedback, collaboration and data-sharing partnerships.

Contact the lab