Estimation of Burned Areas In Forest Fires Using Artificial Neural Networks

M. Hanefi Calp

Karadeniz Technical University

Utku Kose

Suleyman Demirel University

Introduction: This article is the product of the research “Developing an Artificial Neural Network Based Model for Estimating Burned Areas in Forest Fires”, developed at Karadeniz Technical University in the year 2020.

Problem: Forest Fires are an issue that greatly affect human life and the ecological order, leaving long-term issues. It should be estimated because it is not known when, where and how much the fire will be in the area.

Objective: The objective of the research is to use artificial neural networks to estimate the burned areas in forest fires.

Methodology: A feed-forward backpropagation neural network model was used for estimating the burned areas.

Results: We performed a performance evaluation over the proposed model by considering Regression values, Mean Absolute Percentage Error (MAPE) and Mean Square Error (MSE). The results show that the model is efficient in terms of its estimation of burnt areas.

Conclusions: The proposed artificial neural network model has low error rate and high estimation accuracy. It is more effective than traditional methods for estimating burned areas in forests.

Originality: To the best of our knowledge, this is the first time that this real, unique data has been used for building and testing the model’s estimations and the improvements that have been made in producing results faster and more accurately than with traditional methods.

Limitations: Since there are regional differences over different forest areas, effective criteria need to be analysed regarding the target regions.  

Keywords: forest fires, burned areas, artificial neural networks, machine learning, estimation
Published
2020-09-30
Downloads

How to Cite

[1]
M. H. . Calp and U. . Kose, “Estimation of Burned Areas In Forest Fires Using Artificial Neural Networks”, ing. Solidar, vol. 16, no. 3, pp. 1–22, Sep. 2020, doi: 10.16925/2357-6014.2020.03.08.
Metrics
File downloads
505
Twitter
3