We invite you to join the September meeting of the Environment-Vulnerability-Decision-Technology (EVDT) Community. The purpose of the EVDT Community meetings is to share research approaches led by the Space Enabled Research Group and our collaborators that combine insights into social, environmental and policy challenges that can be addressed with geospatial information systems. Learn more about the EVDT Framework here; and read publications from past EVDT projects here.
The meeting will take place on Tuesday, September 2nd from 11:00am to 12:00pm ET. Join the Zoom link here. If you missed the previous meeting, you can find the recording here.
Here is the time-zone information for various locations: Mexico City - 9:00 am to 10:00 am; Rio de Janeiro, Brazil - 12:00 pm to 1:00 pm; Luanda, Angola - 4:00 pm to 5:00 pm; Benin - 4.00 pm to 5.00 pm; Ghana - 3.00 pm to 4.00 pm; Jakarta, Indonesia - 10:00 pm to 11:00 pm; Los Angeles, USA - 8:00 am to 9:00 am.
If you are planning to attend the EVDT Meeting on September 2nd, please fill out this form in advance.
We are also collecting ideas and volunteers for side events (discussions, longer talks, workshops, teaching sessions, etc.). Want to get feedback and advice on some project idea that you have? Want to learn how to use ArcGIS StoryMaps or Google Earth Engine but don't know who to ask?
Whether you want to host an event or want to suggest an event for someone else to host, please fill out this form with your ideas! These can be short, 30 minute events or long multi-hour workshops. The format and topics are very open.
Agenda:
All times are in ET
Introduction | 11:05 AM
Opening remarks from Sharif Islam
Presentation | 11:10 AM
Speaker name: José Bagur
Presentation Title: Edge AI for Autonomous Cloud Detection in Guatemala's Second CubeSat
Abstract: Building on the success of Guatemala's first satellite QUETZAL-1, the Universidad del Valle de Guatemala's Aerospace Laboratory is developing QUETZAL-2, a 2U CubeSat that relies entirely on open-source design principles and commercial off-the-shelf (COTS) components. The satellite's Machine Intelligence for Layer Observation (MILO) payload demonstrates how accessible technologies can enable sophisticated Earth observation capabilities, integrating an OpenMV Cam H7 Plus, an open hardware platform, with a TinyML model developed on Edge Impulse, a free cloud-based platform for creating machine learning models for edge devices. The AI model, trained using publicly available satellite imagery datasets, performs real-time cloud detection to automatically filter and prioritize clear images for transmission, addressing bandwidth limitations inherent in CubeSat communications. MILO also features UVG's first fully in-house developed onboard computer, designed with open-source tools and built entirely from COTS components, which manages the payload operations and interfaces with the satellite's main OBC via standard communication protocols. All hardware designs, software code, AI models, and operational data from QUETZAL-2 will be published under open licenses and made freely available, enabling other institutions worldwide to replicate, modify, or improve upon the system. This presentation will detail MILO's technical architecture using COTS components, the development process leveraging open-source tools and public datasets, integration challenges and solutions, testing methodologies, and how this approach not only advances Guatemala's space capabilities but also contributes to the global knowledge base for cost-effective, AI-enabled small satellite missions.
Organizational Affiliation: University of Valle de Guatemala
Q&A | 11:45 am
Meeting Adjourn | 12:00 pm