Research & Projects

A running log of research threads, software, and applied work — from satellite disease surveillance to AI-driven maritime analytics.

Plaknit — thesis open notebook

Every commit in plaknit captures how I'm building the Satellite Anthrax Surveillance thesis—from PlanetScope ordering and QA to the automation that makes those mosaics repeatable on laptops or HPC schedulers.

PlanetScope mosaic engine

GDAL-powered masking cleans each strip against its UDM raster before Orfeo Toolbox stitches everything together with RAM hints tuned for multi-hundred scene jobs. Documents how UDM-masked strips become seamless mosaics for the Texas Anthrax Triangle.

Landcover classification

The plaknit.classify module handles Random Forest training and inference against Planet stacks for disease-surveillance mapping.

Satellite anthrax surveillance — thesis research

My master's thesis integrates PlanetScope multispectral imagery, cloud computing, and ecological niche modeling to map environmental precursors to anthrax outbreaks in the Texas Anthrax Triangle. The goal is a reproducible remote sensing framework that other researchers can adapt to new anthrax-endemic regions.

  • High-resolution land cover and phenological analysis of outbreak landscapes.
  • Ecological niche modeling with HPC-accelerated workflows.
  • Open-source tooling via plaknit to make the pipeline reproducible.

Maritime surveillance AI/ML — ORNL

During two terms at Oak Ridge National Laboratory's Remote Sensing Group, I joined an interdisciplinary team applying machine learning to maritime domain awareness. The work centered on building scalable pipelines for detecting and characterizing vessels in commercial satellite imagery.

  • Deep learning pipelines — developed and evaluated TensorFlow models for vessel detection, benchmarking output against operational performance thresholds.
  • Sensor assessment — evaluated commercial satellite and sensor modalities for maritime surveillance suitability across resolution, revisit, and coverage dimensions.
  • Analytic automation — built processing automation that reduced manual handling in the imagery-to-intelligence workflow.

Field campaigns

  • Cataloochee Valley UAS survey (2025) — aerial imaging in Great Smoky Mountains National Park supporting NSF RAPID Grant #2501466.
  • Heathsville archaeological survey (2025) — aerial data collection with the UTK Department of Anthropology in Heathsville, Virginia.
  • Hog-spot monitoring (2025) — UAV-based spatial analysis of feral hog activity, presented at TN View Annual Webinar.

Teaching assets

Lab materials I write for UTK courses double as documentation for my own workflows — what I teach is what I run.

Conference talks

  • ASPRS Mid-South (March 2026, Oak Ridge) — Introducing plaknit: An Open-Source Python Software Package for Large-Scale PlanetScope Imagery Processing.
  • AAG Symposium on Spatial AI and Data Science (March 2026, San Francisco) — Mapping High Resolution Anthrax Ecologies: Integrating PlanetScope Data and High-Performance Computing for Ecological Niche Models.
  • AAG Disease Ecologies Session (March 2025, Detroit) — Landscape or Founder Effect: Examining differences in land cover characteristics associated with two distinct lineages of Bacillus anthracis in the Texas Anthrax Triangle.
  • ASPRS Mid-South (April 2025, Oak Ridge) — High-Resolution Land Cover and Phenological Analysis of Anthrax Outbreak Landscapes in the Texas Anthrax Triangle.
  • TN View Annual Webinar (September 2025) — Hog-Spot Analysis: Monitoring the Fiends of the Smokies with UAS and PlanetScope Imagery.
  • UTK Geo-Symposium (January 2025) — Narrowly Avoided the Past: Remembering Nathan Bedford Forrest at Parker's Crossroads Battlefield.

Let's collaborate

If you want to adapt plaknit to a new landscape, discuss maritime or disease-surveillance remote sensing, or need a workshop on PlanetScope processing, I would love to talk.