Comparison of CNN and Vision Transformers for Wildfire Detection: A Proxy for Stubble Burning

Stubble burning is a widely used agricultural practice in India and several other countries that greatly contribute to heavy air pollution. It is done to clear crop residues, usually from April to May and October to November. The gases released during this practice, including particulate matter and greenhouse gases, pose serious health and environmental risks [5]. This research focuses on classifying wildfire burning using satellite imagery as a proxy for stubble burning. We employ deep learning approaches like Convolutional Neural Networks and Vision Transformers to analyze their accuracy, computational efficiency, and real time suitability for environmental monitoring and potentially reduce wildfire burning's impact on air quality.