Aaron Maxwell
West Virginia University, WV·Earth & Environmental Science·Early-career
Humans reshape landscapes through mining, construction and other activities, and natural processes such as landslides keep changing them too. Huge geospatial datasets now record these changes, including satellite images, historic maps and nationwide high-detail elevation data.
- Fields
Earth & Environmental Science
Computer Science
- Topics
geomorphology
deep learning
convolutional neural networks
- Grants
- 1 known grant, through Aug 2027
What they work on
- Develop convolutional neural network methods for semantic and instance segmentation to extract geomorphic features from LiDAR point clouds and terrain derivatives such as hillshades.
- Map historic mining and other human landscape modifications from historic topographic maps and elevation data using deep learning.
- Advance the use of deep learning for geomorphic mapping at regional to global scales, generating datasets that support documentation and modeling of hazards and landscape change.
- Create training and instructional materials in geospatial data science and coding for future and in-service high school STEM teachers and their students.
More about the research
However, pulling useful information from these data efficiently is still difficult. This project asks how deep learning, a type of artificial intelligence, can map these landscape features accurately over large regions.
Why it matters
Accurate, wide-area maps of landscape change could help communities and agencies document and model geomorphic hazards such as landslides, monitor resource extraction such as mining, and quantify human-caused landscape change. Turning large public elevation and map archives into usable datasets could support better-informed land and hazard decisions.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Adriana Trias Blanco
Rowan University, NJ·Engineering
Assess the translation potential of a LiDAR-based bridge inspection platform through industry interviews with transportation agencies, engineering firms and asset managers.
Kristin Sweeney
University of Portland, OR·Earth & Environmental Science
Compare how drainage divides move in soil-mantled landscapes with soil creep versus steep rocky landscapes shaped by landslides, using digital topographic mapping and field sediment measurements.
Majid Bagheri
Savannah State University, GA·Computer Science
Train deep neural networks, convolutional neural networks and long short-term memory networks on water, soil and air datasets to support pollution control, using both training from scratch and transfer learning.
The school
West Virginia University: 15 labs on PI Pages, $259M a year on research, 201 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.