Summary:-Optimist, Software Engineer, Marine Scientist
-PhD Student in Marine Science and Conservation with Dave Johnston in the Marine Robotics and Remote Sensing Center at Duke University
- machine learning approaches to remote sensing analysis
- the intersection of ocean and space exploration
- drones and autonomous systems for marine science and exploration
- machine learning for large-scale environmental monitoring
- adventure sports (paddle boarding, climbing, sailing, surfing)
Work and Focus
I'm currently a PhD student with Dave Johnston
in the Marine Robotics and Remote Sensing Center at the Duke University Marine Laboratory.
My doctoral research focuses on machine learning approaches to remote sensing analysis, coordinating satellite, drone, and in-situ
environmental monitoring, and understanding how autonomous systems will benefit field scientists - all with an emphasis on coastal
and polar environments. I'm particularly interested in systems that will be relevant for marine science as well as for exploring
and understanding other bodies within our solar system.
Before joining the Duke Marine Lab, I worked at Harvard University as a research technician in the Girguis Lab, spent time at Moon Express as a Software Engineer developing their ground data systems and engineering team tools, and served a year and a half as Chief Technology Officer at WayPaver Foundation. I graduated from the University of North Carolina as a Morehead-Cain Scholar with a degree in Computer Science where I started UNC Students for the Exploration and Development of Space, helped start the Sigma Phi Society, and researched computer vision in the Alterovitz Lab.
Open Source Coding Projects
Quick overview of open-source projects.
A curated list of deep learning papers in ecology.
Educational Resources on Neural Networks for Ecology and Remote Sensing.
Summary of recent journal and conference publications.
Integrating Drone Imagery into High Resolution Satellite Remote Sensing Assessments of Estuarine Environments
Patrick C. Gray, Justin T. Ridge, Sarah K. Poulin, Alexander C. Seymour, Amanda M. Schwantes, Jennifer J. Swenson, and David W. Johnston
Published in Remote Sensing - 2018
Satellite imagery now permits dynamic wetland management, yet mapping requires significant fieldwork to run classification algorithms and estuarine environments can be difficult to access. In this study we used drone imagery to create training data to classify WorldView-3 and RapidEye satellite imagery of the Rachel Carson Reserve in North Carolina, USA. We examine change between 2004 and 2017 and conclude that drones can be highly effective in training and validating satellite imagery.
A Convolutional Neural Network for Detecting Sea Turtles in Drone Imagery
Patrick C Gray, Abram B Fleishman, David J Klein, Matthew W McKown, Vanessa S Bézy, Kenneth J Lohmann, David W Johnston
Published in Methods in Ecology and Evolution - 2018
Drones are a potentially promising tool for assessing marine animal populations, but a central challenge lies in analyzing the data generated. Neural networks are emerging as a method for automating detection across domains and can be applied to imagery to generate new population‐level insights. We used a neural network to count turtles in drone imagery and our model detected 8% more turtles than manual counts while effectively reducing the manual validation burden from 2,971,554 to 44,822 image windows.
Drones and Convolutional Neural Networks Facilitate Automated and Accurate Cetacean Species Identification and Photogrammetry
Patrick C Gray, KC Bierlich, Sydney A Mantell, Ari S Friedlaender, Jeremy A Goldbogen, and David W Johnston
In Review in Journal of Applied Ecology
Photogrammetry is contributing to the study of marine mammal individual health, population structure, and understanding changes within communities. This study combines drone surveys with convolutional neural networks for species identification and morphometric assessment of large marine vertebrates. We validate the accuracy of these workflows with traditional approaches. This study presents substantial new morphometric data on humpback, minke, and blue whales and describes an open source CNN for automated species identification and photogrammetric measurements from drone imagery.
Convolutional Neural Networks for Detecting Great Whales from Orbit in Multispectral Satellite Imagery
Patrick C Gray and David W Johnston
IJCAI 2018 - AI for Wildlife Conservation Workshop
Monitoring great whales across their broad spatial scales is prohibitively time consuming using manual analysis methods. We describe a framework using convolutional neural networks to detect great whales in high res satellite imagery. We outline challenges both in data management and model development and propose future research directions for overcoming these obstacles.
Writings and Readings and Talks
My reading history.
In addition to my reading history I've included link to various writings of mine and presentations that may be of interest.
Writings and Docs:
- SEDS SpaceVision - November 2014
- NC Governor's School - July 2015
- New Worlds - October 2015
- USC, Space Studies - December 2015
- TTU SEDS March - 2016
- NewSpace WayPaver Foundation Presentation - June 2016
- NC Governor's School Campus Wide Presentation - Becoming a Spacefaring Species - June 2016
- NC Governor's School Seminar - Lunar Settlement - June 2016
- NC Governor's School Main Presentation - Searching for Life and Spreading our Own - June 2017
- NC Governor's School Seminar/Elective - Exploration as a Mechanism for Survival - June 2017
- UNC SEDS - Vision for Space and Careers within the Industry - November 2017