Key Points
- Sagri Co., Ltd. has developed a real-time disaster damage map that integrates satellite data, AI analysis, and field reports to assist in rescue and relief efforts.
- CEO Shunsuke Tsuboi is requesting citizens in affected areas to submit photos with latitude and longitude (GPS) data via a dedicated LINE official account.
- The map visualizes critical infrastructure status, including road closures, evacuation center locations, and the operational status of medical facilities.
- By prioritizing verified GPS-tagged photos over unverified social media posts, the platform provides high-reliability data for government agencies and rescue teams.
Full Transcript (English)
Guest (Shunsuke Tsuboi): To start with what we need most right now, we want to collect as many photos from the field as possible that include latitude and longitude data. Information about where damage has occurred is currently very fragmented. Information on platforms like X (formerly Twitter) often lacks credibility, so we have set up a system where people can submit data via LINE.
Guest (Shunsuke Tsuboi): If you take a photo and, with your permission, send it to us, the photo along with its latitude and longitude will be sent to our system. We then use a combination of AI and human verification to filter for high-credibility and high-urgency information to display on the map. This type of real-time field data is what we are currently lacking the most.
Guest (Shunsuke Tsuboi): Currently, the map includes data from the Prefectural Police Headquarters’ disaster response center, the Fire and Disaster Management Agency, and major media outlets. We also check government information portals every ten minutes to update the status of evacuation centers and medical facilities. Furthermore, we integrate data from the RTDS (Real-time Disaster Management System) to visualize which roads are currently passable, indicated by the blue lines on the map.
Host (Fuki Kamai): So, by uploading photos taken on a smartphone with location services enabled, real-time information is added to the map?
Guest (Shunsuke Tsuboi): Exactly. Information on social media is often difficult to verify. While we do link to some social media posts, we give them lower priority. We want to populate the map with real, GPS-tagged data. For example, if someone sees a collapsed road and sends a photo, that information can be reflected on the map almost immediately after our verification process. I am also coordinating with local mayors to expand this effort.
Host (Fuki Kamai): Regarding the satellite and drone footage you mentioned, are you manually inputting damage information seen in those images?
Guest (Shunsuke Tsuboi): It depends on the image quality. We are moving forward with AI analysis, but the first step is simply getting the images onto the map as a layer. We are currently working with partners like Terra Labo to integrate drone footage and satellite providers for the latest imagery. This is updated continuously.
Host (Fuki Kamai): You mentioned the RTDS data for roads. How does that look on the map for someone trying to navigate, for example, from an airport to a specific town?
Guest (Shunsuke Tsuboi): The map shows road segments, but it isn’t perfect yet. Some data cannot be shared directly due to regulations, so we update it manually where necessary. This is why ground-level information is so vital. Official open data on specific road damage often isn’t available immediately, so crowdsourced photos of cracks or blockages help fill those gaps. Without high-credibility data, the map becomes confusing, so we focus on verified reports.
Host (Fuki Kamai): Capturing the damage to private homes is also important for future disaster subsidies and for local governments to understand the scale of the disaster, isn’t it?
Guest (Shunsuke Tsuboi): Yes. While local governments collect this data, they often don’t have the means to visualize it on a digital map in real-time. We built this tool quickly to bridge that gap between citizens, rescue teams, and the administration.
Host (Fuki Kamai): What was the primary motivation for Sagri to start this project now?
Guest (Shunsuke Tsuboi): During the Noto Peninsula earthquake, we realized that satellite analysis alone has significant delays. There was a clear need for a tool that visualizes real-time field data. With current AI technology, we can implement these tools rapidly. Our goal is to ensure that everyone—from rescue workers to affected residents—has access to the same reliable information so that support can be directed where it is needed most.
Guest (Shunsuke Tsuboi): Finally, I want to reiterate: please send photos with GPS data to our LINE account. We will verify and add them to the map to help local authorities and the public stay informed. If there are other features or types of information you need, please let us know through the same channel.
