"Cooking vs. Fire Detection with Beep in 10 Seconds"…Development of Intelligent Fire Detector

Korea Research Institute of Standards and Science Accurately Detects Small Flames Within 3cm
Uses Artificial Intelligence to Distinguish Flames with People as 'Cooking' Flames

Fire. Not related to the article.

[Asia Economy Reporter Kim Bong-su] An intelligent fire detector that distinguishes between flames used for cooking and actual fires, and sends an alarm within 10 seconds only in the case of a real fire, has been developed.

The Korea Research Institute of Standards and Science (KRISS) announced on the 30th that researchers from the Nondestructive Evaluation Team at the Safety Measurement Research Institute have developed an intelligent fire detector that distinguishes between flames used by people and actual fire flames, recognizing only flames caused by real fires and notifying within 10 seconds of ignition.

The institute explained that it can accurately detect small flames as small as 3 cm occurring at the very early stage of a fire, which is expected to greatly help in early fire suppression and evacuation.

70% to 80% of fire fatalities are caused by suffocation due to smoke and toxic gases. To prevent fires and avoid casualties, it is very important to detect and notify of fires at an early stage. However, existing fire detectors can only detect fires at the second stage, which is one minute after the initial ignition, making suppression and evacuation difficult due to smoke and flames. Sprinklers also only activate when the indoor temperature exceeds 72 degrees Celsius, so they cannot ensure the safety of people inside.

The newly developed intelligent fire detector recognizes fires at the first stage, which corresponds to the very early stage of a fire, and notifies through its own alarm and a smartphone app. It can confirm the coordinates of the flame’s location, enabling automatic localized extinguishing when linked with fire suppression devices. Notably, the false alarm rate is below 3%, which is highly reliable compared to the 34% to 50% false alarm rates of existing smoke or heat detectors.

The research team improved flame recognition by introducing fusion sensing technology that combines infrared sensors and infrared thermal imaging sensors. They utilized the advantage of infrared sensors to quickly detect flames by using specific CO2 wavelength bands of the flame. In particular, they developed and applied an artificial intelligence technique that tracks people within the space where the fire detector is installed to recognize that flames used for cooking or work are not fires. The algorithm simultaneously recognizes people and flames with an infrared thermal imaging camera and distinguishes whether the flame is used by a person or is a fire based on whether both a person and flame coexist within the sensing space.

Choi Man-yong, lead researcher at KRISS’s Safety Measurement Research Institute, said, “This is a fire monitoring technology that improves the shortcomings of existing fire detectors to quickly recognize only actual fire flames.” He added, “It will increase the reliability of fire alarms, alleviate public anxiety, and help with early fire suppression.”

This content was produced with the assistance of AI translation services.