Fire and smoke detection system with alarm using arduino.

This survey paper reviews the challenges and recent advancements in IOT smoke and fire detection systems, with particular focus on both indoor and outdoor environments. The main problem addressed is the high false alarm rates (ranging from 3.4% to 29.49% across various systems) and the challenges posed by environmental variability, dataset scarcity, and the complexity of real-time detection. The paper critically examines key methodologies, including traditional approaches, hybrid methods, and domain transfer-based tools, highlighting their evolution and current trends. This survey also provides an in-depth analysis of publicly available datasets and evaluation metrics, such as detection accuracy (ranging from 79.66% to 98.72%), robustness to dynamic environments, and real-time processing capabilities (with some systems achieving up to 333 frames per second (FPS). By synthesizing insights from 33 papers published between 2013 and 2024, the survey not only summarizes the current state of the art but also identifies emerging trends, such as the increasing use of automatic feature learning and multi-fusion systems, which have demonstrated significant improvements in detection accuracy. The paper concludes by advocating for future research focused on improving system robustness and reducing false alarms through the integration of visible range cameras and traditional sensors, with the goal of achieving more accurate and reliable fire detection in surveillance systems.