Projects at ORNL
Data set creation for forest canopy derived human decomposition signatures
I lead a collaboration with the University of Tennessee to create a preliminary dataset of forest canopies in the presence of human decomposition to determine if we could detect human decomposition from multispectral data. We collected multispectral data from forest canopies via unmanned aerial systems (UAS) to determine the estimated nitrogen content through vegetation indices in order to compare the nitrogen content between areas with decomposition and control areas.
FSSR: Forest Sensing for Search and Recovery
I lead the research and development of a computer vision model for the classification of potential human decomposition signatures from forest canopy. The model was trained on data collected by ORNL at UTK’s Anthropology Research Facility using the YOLOv8 algorithm (Ultralytics).
eDNA Bot
Assisting with the research and development of an autonomous environmental DNA collection device. Project on going.
Terrestrial LiDAR Scanning (TLS) data pipeline
Assisting in the creation of a TLS data pipeline for utilization in forest ecology, carbon capture, biomass estimation, and many other forest health metrics. The goal of this research is to be able to create a TLS pipeline that can process raw data and create quantitative structure models, allometry metrics, leaf area index, leaf angle, and architecture of specific trees.
Metagenomic Analysis of Populus trichoderma and Populus deltoides
Metagenomic analysis of two Populus species, trichoderma and deltodies across different seasons. We are investigating genetic diversity and microbial ecology across time.
Image Analysis for Plant~Microbe Interactions
Fungal~Plant VOC Interactions
Utilizing artificial intelligence algorithms to analyze different petri dish experiments investigating how different environmentally relevant chemicals interact with Populus associated microbes.
Dark Septate Endophyte Metabolome
Running data analysis on various VOC data collected from fungal-plant interactions. Looking at both plants and fungi in a soil mesocosum as well as Populus in a petri dish with various fungi. These VOCs are being compared to see which ones are present under different conditions.
UNET Architecture for Plant Phenotyping
Completed Projects:
Investigating the metabolome of Hyaloscypha finlandica and other dark spetate endophytes using a combination of metabolomics and computational genomic analysis to understand what metabolites are being produced and what biosynthetic gene clusters could be responsible for it.
Building a UNET machine learning algorithm for image segmentation and analysis of plant images that come from the Advanced Plant Phenotyping Lab (APPL).