Waving: From Space to Ocean (Senior Project)

Collaborated with NASA Goddard Program, University of Maryland, and UNT

Jan 2024 – May 2025
01

Problem

NASA's PACE mission produces large-scale, high-dimensional oceanographic and atmospheric satellite data that is difficult for non-experts to interpret or interact with. The goal was to transform raw satellite data into a real-time interactive system that enables scientists, educators, and general audiences to explore complex datasets through intuitive gesture-based interfaces — without requiring domain expertise.

02

Constraints

  • Real-time performance requirements with multi-user simultaneous interaction
  • High-throughput stereo camera input (ZED SDK) requiring low-latency processing
  • Deployment on constrained exhibit hardware with no dedicated GPU infrastructure
  • Robustness across variable physical environments — lighting, spacing, and calibration
  • Non-technical end users: scientists, educators, and general public audiences
03

Approach

Designed a modular pipeline combining computer vision, real-time rendering, and gesture recognition. Built gesture interaction using ZED stereo cameras and OpenCV with CUDA acceleration for tracking and interpreting user movement. Developed a configurable ImGui-based setup wizard to automate deployment, calibration, and system validation, reducing friction for non-technical operators. Integrated the full pipeline into Unity to render interactive PACE satellite visualizations across multiple display environments.

04

Architecture

05

Metrics

MetricBaselineAchieved
Setup & calibration timeManual multi-step process~70% reduction via automated wizard
Gesture recognition latencyNot real-time<100ms end-to-end
Simultaneous users supportedSingle user2–4 participants
Runtime errorsManual validation onlyEliminated via automated pre-flight checks
06

Product Impact

Delivered an interactive exhibit system used to demonstrate NASA PACE mission data for public education and scientific outreach. Enabled intuitive exploration of satellite-derived oceanographic data without requiring domain expertise, bridging the gap between raw scientific output and accessible visualization. Improved deployment scalability so non-technical staff could independently configure and run the system. Currently exhibited at the Kennedy Space Center Museum, Washington D.C., with a scheduled national museum tour.

07

Tech Stack

C#
C++
Python
Unity
OpenCV
ZED SDK
CUDA
ImGui
OpenGL
React
TypeScript
GraphQL
08

Links