Finding victims in post-disaster spaces

When earthquakes strike, people often get trapped in buildings. Search and rescue teams can pinpoint some victims using sniffer dogs and sensors. But a new European system that takes pictures during or after a building collapse promises to save many more lives.

The ‘low-cost catastrophic event capturing system’ is the fruit of the IST Loccatec project. It comprises capturing devices (small, autonomous infrared cameras) and a portable central unit for the rescue teams. The cameras – pre-installed in a building – only record images when they detect a structural collapse. Nearby rescue teams can wirelessly download these images and even set up an audio-link with the devices. The priority is to find where people are and the best routes to reach them.

Early tests were done in 2002, in the Italian town of San Juliano De Pulia – scene of the terrible earthquake that killed many schoolchildren. “After the disaster, we placed our cameras inside several damaged buildings earmarked for demolition,” says Uberto Delprato, project coordinator. The resulting eight films, recorded during demolition, were used to fine-tune the system’s ability to trigger image-recording during a building collapse.

The software algorithms for this triggering mechanism are unique. They can identify building collapse through a significant change in monitored structural elements – such as a when a roof falls. Loccatec has applied for a European patent for this software, and the hardware that stores and wirelessly transmits images.

While the system is similar to a burglar alarm, there are differences. “A burglar alarm continuously watches for break-ins, but does little else,” notes Delprato. “Loccatec only starts recording after the onset of a building collapse. And these are not just standard cameras, because they include lots of software.” Also unique is the system’s ability to multihop, allowing one device to act as a network bridge to others.

The images are stored on a chip at one frame per second, a rate which increases when collapse is imminent. A typical installation would have four cameras for a meeting room and ten for an auditorium. The cameras are placed on naturally strong elements, such as pillars. It is the task of the system’s algorithms to decide what is impending disaster and what is just human movement.

Public buildings, especially schools and hospitals, would be the first market for Loccatec. Not so much because of the cost of installation, but the need for someone to be responsible for the system.

One challenge is convincing the public that these capture devices are not secret cameras. The project coordinator is reassuring: “No images are recorded unless there is a disaster detected. Moreover, the transmission of images to a central unit is protected by a password and special protocol.”

In late 2004, the project will test the prototype system in Greece, using shaking tables to mimic an earthquake. Says Delprato: “Commercialisation of the system is a distinct possibility, provided a company takes it up. Meanwhile, we are going for the patent and want to exploit the system as a group of project partners.’

He foresees other applications for a low-cost device with a triggering mechanism. It could be used to mitigate disasters in chemical plants or to detect a bridge collapse during a flood.