EO in Orbit: Scientific webinars – AI and climate change apps with EO data
- Type: Webinar
- Theme: Earth observation (EO) capabilities, artificial intelligence (AI) and climate change apps
- Time: 11:00 a.m. to noon ET
- Duration: 1 hour
- Cost: Free
- Location: Virtual
- Language: English
- Target audience: Industry, academic institutions, media, not for-profit organizations, scientists, government.
Presentations will focus on the use of AI to enhance EO capabilities and the development of climate change-focused apps.
- Augmenting Earth Observation Capabilities in Orbit Enabled by Artificial Intelligence on the Edge
- Notes from the Field: Building Climate Change Apps with EO Data – Lessons Learned
Augmenting Earth Observation Capabilities in Orbit Enabled by Artificial Intelligence on the Edge
Presentation 1 (in English)
From 11:00 to 11:30 a.m.
As EO becomes more accessible and affordable with the proliferation of satellites and lower launch costs, there is a growing demand for enabling autonomy for EO data processing and analysis for complex missions in communications-denied environments. EO systems can increasingly leverage AI to enable intelligent onboard data analysis and speed up decision-making processes such as intelligent prioritization of data to downlink to Earth. With novel technology to deploy and maintain the use of AI in spaceflight, Mission Control is pioneering how EO missions can leverage onboard AI to achieve several benefits such as:
- reducing downlink times for pertinent information and therefore reducing overall "
- enabling faster responses to time-critical events such as natural disasters and other emergency scenarios
- reducing cognitive burden on operators and other ground segment resources by providing recommended courses of action, automated workflows, and shifting data processing from ground segment computers to flight computers
- reduced spacecraft downtime by enabling intelligent onboard fault detection and prediction, diagnostics and recovery processes
In this presentation, we provide an overview of benefits to leveraging AI capabilities at the edge for EO operations, as well as an overview of barriers to adoption that range from lack of established verification and validation processes to flight hardware constraints. We then introduce Mission Control's Spacefarer AI framework designed to address several challenges. Spacefarer AI is being developed to assist data scientists and software developers in their goals to design, develop, deploy and run AI models on flight computers that can benefit their EO operations.
Notes from the Field: Building Climate Change Apps with EO Data – Lessons Learned
Presentation 2 (in English)
From 11:30 a.m. to noon
Climate change-focused software projects are challenging – and very different from regular software development. They rely on a complex combination of data sets, including imagery and observations taken by Earth-observing satellites. They impose many significant non-technical as well as technical constraints and hurdles that are not regularly faced in typical software delivery. They often require a deep level of scientific or technological understanding to produce a useful result. They impose a steep learning curve on software development teams. Successful project outcomes generally require sustained support and input from numerous stakeholders including government, academia, private sector, community groups and even individuals. They are risky and expensive.
This presentation identifies some of the key challenges, strategies, and lessons we have learned while building climate change-focused software applications that use space-based Earth observation (including as part of the Canadian Space Agency smartEarth initiative). We will first briefly identify major architecture, design, development and delivery challenges of a climate change-focused project that uses satellite imagery. Then we will identify practical strategies and approaches we took to overcome these. No prior technical knowledge from the audience is assumed or required.
- Date modified: