Leveraging Machine Learning for Rapid Response in Wildfire Management
Luis Miguel Velazquez Rodriguez
Co-Presenters: Individual Presentation
College: The Dorothy and George Hennings College of Science, Mathematics and Technology
Major: Computer Science
Faculty Research Mentor: Daehan Kwak
Abstract:
Title: Leveraging Machine Learning for Rapid Response in Wildfire ManagementAuthor: Luis Miguel Velazquez Rodriguez, Department of Computer Science and Technology, Kean UniversityA wildfire is described as an unwanted and uncontrolled burning of a natural area, such as forests or grasslands. Throughout history, various methods have been utilized to detect and prevent wildfires in a timely manner. However, with the rise in global warming, wildfires have become more frequent, highlighting the need for new and innovative solutions. Research from the Center for Climate and Energy Solutions indicates that a 1-degree Celsius rise in temperature could result in a median annual burned area increase of up to 600%.The goal of this research is to show how machine learning can help with the detection of wildfires through the use of developing technology. The research will implement machine learning through the use of convolutional neural networks (CNNs), long short-term memory networks (LSTMs), and the Random Forests algorithm. By leveraging the power of CNNs, LSTMs, and Random Forests, this research aims to improve the early detection of wildfires, ultimately providing faster response times and reducing the devastating impact of these disasters on both human life and the environment.Two datasets were used and combined into one dataset containing 500 images. The images were divided into training, testing, and validation sets. The training set had 200 images, and the test and validation sets had 50 images each. Each set was split evenly into 'fire' and 'no fire' labels. In addition to the image data, temporal information, such as weather patterns over time, was incorporated using LSTMs to analyze the progression of environmental changes. Furthermore, environmental factors such as temperature and wind speed were included in the model.As technology continues to evolve, integrating machine learning models with real-time data could significantly enhance early detection and response systems. The research aims to enhance wildfire prediction strategies, advancing the potential of machine learning in environmental monitoring.