ANDRES HERNANDEZ-SERNA
Scientist
Scientist
“We are at the very beginning of time for the human race. It is not unreasonable that we grapple with problems. But there are tens of thousands of years in the future. Our responsibility is to do what we can, learn what we can, improve the solutions, and pass them on" Richard Feynman. 1918 - 1988
RESEARCH
I'm a Principal Faculty Specialist (100% research) in the Global Land Analysis and Discovery (GLAD) group at the University of Maryland’s Department of Geographical Sciences. I also serve as a University Senator, and an Advisor to the Artificial Intelligence Interdisciplinary Institute (AIM). My work supports GLAD’s core mission: using Earth observation imagery to monitor and understand global land cover change.
My research applies deep learning and high-performance computing (HPC) to large-scale remote sensing. I work with petabyte-scale geospatial data LiDAR, high-resolution imagery, and time series to analyze landscape change. I process data from major NASA missions, including Landsat and GEDI, to produce actionable, long-term insights into global land dynamics.
Key areas of my research include:
Deep Learning in Spatial Data Analysis: I spearhead the development and implementation of deep learning frameworks for analyzing spatial data. My work involves leveraging U-Net architectures and Generative Adversarial Networks (GANs), along with a 24-GPU setup, to significantly enhance the accuracy and efficiency of satellite image analysis.
NASA LiDAR Data Processing: I co-developed pyGEDI, a Python package for analyzing data from NASA's Global Ecosystem Dynamics Investigation. Additionally, I utilize LiDAR-RIEGL data from drones to assess biomass and tree height in various ecosystems.
Remote Sensing Techniques: I develop algorithms to process and analyze satellite imagery, notably from NASA's Landsat missions and the Planet constellation. My aim is to create detailed time-series maps of urban areas, managing vast datasets with spatial resolutions as 3 meters and time series extending back to 1984 at 30 meters resolution.
Comparative Analysis of Remote Sensing Data: I conduct evaluations on the effectiveness of combining Planet, Landsat, and GEDI datasets for tree height measurement, covering a period from 2019 to the present.
Field Studies and Data Validation: I participate in fieldwork in diverse regions, including South America, Africa, and the US, to validate remote sensing data, especially in relation to agricultural expansion and forest change.
With the support of HPC systems, my research leads advances in spatial data analysis to better understand land use and land cover change. These tools make it possible to manage and analyze complex, large-scale datasets in ways that were not possible before. This is key for making informed decisions in land management, conservation, and sustainable development.