• Kola Harsha Vardhan
• V. Neelima Devi

Forest fires can cause significant damage to the environment and ecosystems, making early detection crucial for prevention and mitigation. In this research, a novel digital image processing approach called Multi-Feature Best Decision-based Forest Fire Detection (MF-BD-FFD) is proposed. The method uses AI-based computer vision techniques to detect fires and smoke from images. Deep learning algorithms, specifically Convolutional Neural Networks (CNN), are employed with image processing and greyscale techniques for fire detection. Once a fire is detected, an alert is sent to the forest control team along with the location information. To enhance the sensitivity of detection, colour and texture features are incorporated, and hybrid decision-making algorithms such as Artificial Neural Networks (ANN), Support Vector Machine (SVM), and K-Nearest Classifier (KNN) are utilized. The output from these algorithms is optimized to select the best decision. Compared to conventional techniques, the proposed method shows a 5% increase in accuracy, improving the overall performance of the system. The automatic identification of forest fires using AI-based computer vision techniques has significant potential in minimizing fire disasters and aiding decision-makers in planning mitigation strategies and extinguishing tactics. The use of multi-feature extraction and hybrid decision-making algorithms can enhance the accuracy and sensitivity of forest fire detection, enabling early intervention and prevention of major damages.

Forest Fire Detection (FFD), Artificial Neural Networks (ANN), Support Vector Machine (SVM), k- nearest Classifier (KNN), Digital Image Processing (DIP)

"Detecting Forest Fire Using Deep Learning Techniques", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 4, page no.f153-f162, April-2023, Available :http://www.jetir.org/papers/JETIR2304521.pdf

"Detecting Forest Fire Using Deep Learning Techniques", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 4, page no. ppf153-f162, April-2023, Available at : http://www.jetir.org/papers/JETIR2304521.pdf