A multimodel fire detection and alarm system for automobiles using internet of things and machine learning

Fire accidents in vehicles lead to the loss of human lives. The fire detection and alarm systems are often error-prone and respond to nonactual indications of fire presence, known as false alarms. The proposed model detects the fire at the smoldering stage and buzzes an alarm if an actual fire or smoke is detected. This system can achieve this real alarm using multiple internet of things-based sensors, namely smoke/gas, flame, temperature, and a visualization camera. The visualization camera continuously captures images of the vehicle to check the existence of fire. Machine learning algorithms are executed on the sensor and image dataset to reduce false alarms and achieve high accuracy of results by using various performance metrics.