Researchers at the Multimedia University (MMU) in Malaysia have developed an omnidirectional thermal visualizer that provides a 360-degree view of the area under surveillance. The device, the researchers say, can monitor more than one lane at a time in airports and crowded areas, making it easier for authorities to quickly identify people with flu or SARS-like symptoms.
Typically thermal imaging cameras deployed in places like airports require one camera per lane to detect SARS or flu-infected people. These cameras can cost upwards of US$30,000; even inexpensive models cost around US $6,000. In order to monitor a larger area at low cost, Dr Wong Wai Kit and his team at MMU developed the omnidirectional thermal visualizer. Consisting of an infrared camera, a customized hyperbolic hot mirror and special algorithms that perform the image processing, the visualizer could, the team reports, allow authorities to reduce both the number of cameras deployed in crowded areas and the overall cost of hardware needed.
UPGRADE TO NEW ATLAS PLUS
More than 1,200 New Atlas Plus subscribers directly support our journalism, and get access to our premium ad-free site and email newsletter. Join them for just US$19 a year.UPGRADE
"Conventional methods require one infrared camera to cover 90 degrees," Kit tells Gizmag. "To get an omnidirectional (360-degree) image, you'd need four infrared cameras, four image processing units, and synchronization circuitry to stitch all the images together. With our invention you only need one infrared camera, an image processing unit, and a hyperbolic hot mirror. There's no synchronization required since a snapshot is an omnidirectional image."
To get an omnidirectional image, the researchers mount the hyperbolic hot mirror on a specially designed camera-mirror holder, so that it's precisely placed against the infrared camera. The mirror is coated with an infrared reflector material, with a reflective index that can be used to get the temperatures of the people being monitored. "It actually reflects the temperature of the person back," explains Kit.
The mirror is able to able to expand the camera's vision from 90-120 degrees to 360 degrees. The omnidirectional image obtained is unwarped with special algorithms, to provide observers with a panoramic image offering a complete wide angle view. The algorithms detect the presence of people, pick out their foreheads, and trace out their body temperatures. Currently the system requires a human operator to look at the screen and pick out people who might have infections; the researchers plan to incorporate an auto-detection feature later on.
"Once the system detects a flu-infected person in a crowded area, we can use CCTV to further pinpoint that person and get a clear image," Kit tells us. "Security and medical personnel can then approach the person for a detailed checkup/further advice to get medical treatment if necessary."
The team is presently working with Telekom Malaysia to enable realtime streaming of captured videos from a remote location with the aid of their Unifi/HSBB broadband services. The goal is to support the government in developing a smart community where the system can be used to stop an outbreak of infectious disease in airports, shopping malls, schools, and hospitals. It's also ideal for monitoring traffic at cross junctions, according to the team, as it can see all four ways. While the device is currently capable of monitoring an area with a radius of 5 m (16 ft), the team aims to increase its range to a radius of 100 m (320 ft) or more.
"We want it to be able to track moving vehicles," Kit tells Gizmag. "The Lahad Datu intrusion incident inspires us to seek a way to apply this system for homeland defense."
The team plans to improve the imaging distance and develop object recognition algorithms to detect aircraft, ships, and all types of vehicles. They also plan to adapt it to detect burglaries, crimes, and incidents such as a person hanging around a car for a long time, fainting, or becoming unconscious. "The system can present object shapes based on their temperature, identify moving targets, and apply pattern recognition, which it can use to identify human behavior," explains Kit.
The researchers are currently drafting a research grant proposal with the Malaysian government to work on a planned metropolitan surveillance system. The device was recently exhibited at the 25th International Innovation and Technology Exhibition (ITEX) at the Kuala Lumpur Convention Center in Malaysia
Source : MMUView gallery - 5 images