Change detection for geospatial imagery, Innovate4Cities hackathon, & GeoAI for investigative journalism
Welcome to GeoAI Horizons - your monthly look at the intersection of geospatial data & artificial intelligence!
Our first edition of GeoAI Horizons discusses change detection for geospatial imagery made easy, how to participate in an AI x City climate action hackathon, and how geoAI can be used in investigative journalism. Let's dive in!
What is change detection & how can it be implemented?
Change detection is a key competency for many industries, including environmental monitoring, urban planning, and infrastructure maintenance. Such functionality allows you to identify changes in your geospatial images over time, such as new construction, deforestation, or other critical environmental shifts.
Change detection might seem straightforward at first glance, often synonymous with “something that has changed.” Semantically, it appears simple, but from a technical perspective, it unfolds into a realm of possibilities, each defining a potential solution in its own right.
Recently, we wrote a guide that offers a straightforward introduction to the topic while providing specific methods and approaches for those looking to deepen their understanding. It can be downloaded for free here.
We also introduced a dedicated automated UI/UX for building change detection datasets and a dedicated change detection machine learning architecture. You can see it in action below and learn more here.
Innovate4Cities hackathon
Can you develop a methodology or process flow to assess the risks and vulnerabilities of any Brazilian city to heat extremes, droughts, sea level rise, floods, and other climate hazards?
That's the challenge set by the AI x City climate action hackathon, organized by the Global Covenant of Mayors for Climate & Energy (GCoM) and partners.
We're thrilled to partner with the hackathon, providing platform access and expert training to help cities build resilience and adapt to climate change. Pitch your solution virtually or in person at the 2024 Innovate4Cities Conference (I4C24) in Montréal, Canada, and receive support to see your idea come to life!
Uncovering ground truth with geoAI for investigative journalism
One notable use case of geoAI is supporting visual investigation teams at news and media organizations. GeoAI can be used in cases involving conflict, environmental issues, human rights, socioeconomic disparities, cultural heritage, urban development, and more.
Earlier this year, we were humbled to learn that an AI-driven investigation by 纽约时报 using Picterra shared in a prestigious Pulitzer Prize. The staff of 纽约时报 won the?2024?Pulitzer Prize?in?International Reporting for "its wide-ranging and revelatory coverage of Hamas’ lethal attack in southern Israel on October 7, Israel’s intelligence failures and the Israeli military’s sweeping, deadly response in Gaza."
As part of its coverage, aerial imagery and Picterra were used to detect bomb craters that showed that one of Israel’s biggest bombs was used routinely in south Gaza. We were particularly impressed by the team's approach to building and iterating the model using Planet satellite imagery and setting up a post-processing workflow to verify detections.
More recently, the investigative research team at Süddeutsche Zeitung used satellite imagery and Picterra to detect and analyze tanks within Russian military bases during the ongoing war in Ukraine.
We're proud to continue to support evidence-based investigative journalism, uncovering ground truth in a wide range of use cases. We are currently working with several international investigation teams on upcoming stories. Feel free to reach out if you need our support.
That's it for our inaugural edition of GeoAI Horizons. Be sure to subscribe so you don't miss future editions! Do you have feedback or something you'd like us to cover? Don't hesitate to get in touch.