<![CDATA[Fire is a disaster that can cause significant losses if not detected early. This study aims to develop a fire detection application based on digital image processing using a combination of color segmentation, texture feature extraction, and machine learning classification in MATLAB. Images in JPG and PNG format are processed by converting the RGB color space to HSV, followed by masking the fire-colored regions for segmentation. The segmented areas are then processed for texture feature extraction using the Gray Level Co-occurrence Matrix (GLCM) method, focusing on four main parameters: contrast, correlation, energy, and homogeneity. The extracted features are used as input for classification using the Random Forest algorithm with 100 decision trees. The system output is in the form of labeled images indicating either “FIRE DETECTED” or “NO FIRE.” The results show that the combination of HSV color segmentation, GLCM texture features, and Random Forest classification has strong potential as an effective approach for fire detection based on digital image.]]>
Copyrights © 2025
