Atlas Overlay — NASA SpaceApps Challenge
Construction planning tool leveraging NASA satellite imagery, terrain segmentation, and CV-based analysis. Awarded 1st Place — Galactic Impact at NASA SpaceApps Cleveland 2021 and Best Trine Research Project (Engineering Multi-Departmental) at Trine University's STEM Symposium.
What this project proves
Prototype builder across CV and product framing
Award-winning terrain and construction-planning prototype using NASA imagery and computer vision analysis.
Core challenge
Use remote sensing and CV to make construction planning more actionable.
Evaluation lens
Prototype design, terrain analysis, and end-user framing.
An award-winning applied CV prototype with clear decision-support value.
Overview
Atlas Overlay is a construction planning tool that combines NASA satellite imagery with terrain segmentation and CV-based site analysis to give pre-build engineering teams a richer automated view of candidate construction sites. It was submitted to the NASA SpaceApps Challenge 2021 in Cleveland and later recognized at Trine University’s STEM Symposium.
This project is best read as an early product-and-prototype signal: taking remote-sensing data and shaping it into a tool that helps planners make better site decisions before construction begins.
Awards
- 1st Place — Galactic Impact, NASA SpaceApps Challenge (Cleveland, OH, 2021)
- Best Trine Research Project — Engineering Multi-Departmental, Trine University STEM Symposium
What I Owned
Led a cross-disciplinary team across computer science, civil engineering, and geosciences. I owned the technical architecture, the computer-vision and terrain-analysis pipeline, and the demo integration that turned the analysis into a usable presentation artifact.
Hard Problems Solved
- Translate raw imagery into planning signal: remote-sensing inputs needed to become something more actionable than a map layer
- Bridge disciplines: the project had to connect CV analysis with civil-engineering and site-planning concerns
- Deliver a usable demo quickly: hackathon-style timelines reward fast iteration, but the result still needed a coherent technical story
Key Features
- NASA Satellite Imagery Ingest: Pulled multi-band satellite tiles for any candidate site and aligned them with local terrain data.
- Terrain Segmentation: CV-based analysis of overlay gradients, slope, and surface composition to flag risk zones.
- ArcGIS Visualization: Layered the segmentation outputs onto interactive ArcGIS maps for site planners.
- Autodesk Revit Integration: Generated realistic building data and structures from the overlay analysis to support reliable pre-planning.
Why It Matters
Atlas Overlay shows that I can frame computer-vision work around a real decision-support use case, not just a model demo, and communicate it well enough to win in both hackathon and academic presentation settings.
Tech Stack
- GIS: NASA satellite imagery datasets, ArcGIS
- CAD: Autodesk Revit (parametric building generation)
- Computer Vision: Terrain segmentation, gradient analysis
- Tooling: Python, image processing libraries