eliasm56
PhD student in Natural Resources & the Environment at the University of Connecticut.
University of ConnecticutStorrs, CT, USA
Pinned Repositories
Arctic-Infrastructure-Detection-Paper
A full deep learning pipeline for satellite image semantic segmentation (originally designed for detection of built structures) using Segmentation Models PyTorch.
IWP-Network-Extraction
Extracting the network of troughs surrounding ice wedges detected by the MAPLE workflow using a simple geospatial vector data processing workflow
Lightweight-U-Net-for-Satellite-Image-Segmentation
An approach to efficient satellite image semantic segmentation with a vanilla U-Net built in Keras, aided by image augmentation and mathematical morphology.
orthogonalize-polygon
Orthogonalize polygon in python by making all its angles 90 or 180 deg
UIMG2DSM
U-IMG2DSM: Unpaired Simulation of Digital Surface Models with Generative Adversarial Networks
HABITAT
HABITAT = High-resolution Arctic Built Infrastructure and Terrain Analysis Tool. A fully-automated, end-to-end geospatial deep learning pipeline designed to map Arctic built infrastructure from <1 m spatial resolution Maxar satellite imagery.
IW-Network-Extraction
Extracting the network of troughs surrounding ice wedges detected by the MAPLE workflow using a simple geospatial vector data processing workflow
pdg-portal
Design and mockup documents for the PDG portal
GeoPandas-IWP-Hex-Summation
Summarize count of ice-wedge polygons within hexagonal grid cells with a CPU-parallelized GeoPandas-based algorithm.
IWP-Data-Retriever
Retrieve shapefiles of IWPs predicted by MAPLE that are requested by a data user.
eliasm56's Repositories
eliasm56/Arctic-Infrastructure-Detection-Paper
A full deep learning pipeline for satellite image semantic segmentation (originally designed for detection of built structures) using Segmentation Models PyTorch.
eliasm56/Lightweight-U-Net-for-Satellite-Image-Segmentation
An approach to efficient satellite image semantic segmentation with a vanilla U-Net built in Keras, aided by image augmentation and mathematical morphology.