Exploring the Potential of Machine Learning Approaches in Fire Detection: A Case Study in Marmara and Kocaeli Region

Usta A., AKBULUT O.

6th International Congress on Human-Computer Interaction, Optimization and Robotic Applications, HORA 2024, İstanbul, Türkiye, 23 - 25 Mayıs 2024, (Tam Metin Bildiri) identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/hora61326.2024.10550458
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Anahtar Kelimeler: classification, decision tree, fire detection, MODIS, random forest, remote sensing, support vector machine, VIIRS
  • Kocaeli Üniversitesi Adresli: Evet

Özet

Fires pose a risk that seriously threatens our daily lives. Nowadays, early fire detection has become important to minimize the damage caused by fires. In this paper, the performances of machine learning approaches in fire detection have been examined using MODIS and VIIRS remote sensing datasets. The experiments have been carried out in Kocaeli and Marmara region. Experimental results show that learning-based approaches can enable classification performance up to 97% in fire detection.