A RULE-BASED FIRE DETECTION FRAMEWORK USING IMAGE PROCESSING FOR REDUCING FALSE ALARMS IN FOREST AREAS

Abstract

Forest fires pose significant threats to ecological systems, property, and human life, necessitating prompt and accurate detection methods. This research presents the design and implementation of a fire detection algorithm using image processing techniques to enhance the verification of fire events in forest areas. The proposed system aims to improve early fire detection accuracy and minimize false alarms, responding to alerts generated from Monitoring Centers (MCs). The study emphasizes the advantages of image-based monitoring systems, including cost-effectiveness, reliability, and ease of integration with existing technologies. The fire verification algorithm follows a structured, five-phase approach: image acquisition, pre-processing, RGB extraction, color space transformation to YCbCr, and rule-based classification. Image pre-processing enhances pixel intensity to better isolate fire-relevant color features. The RGB color components are then extracted and transformed into the YCbCr model to effectively separate luminance from chrominance, a critical step in distinguishing actual fire events from visually similar non-fire objects. Rule-based classification is applied in the final phase to identify and categorize fire and non-fire images with high precision. The model is designed to handle challenging conditions such as smoke interference, which often distorts visual data and leads to false detections. By addressing these challenges, the algorithm ensures improved accuracy in fire event confirmation. The study demonstrates that incorporating image processing methods into fire detection frameworks significantly enhances verification capabilities, contributing to more effective disaster prevention and response strategies in forest environments.