Student: Anshul Sadh Gauri
Advisor: Jaime Dávila, PhD
Committee Member: Erik Learned-Miller, PhD
This thesis focuses on designing and implementing a semi-automated system to detect deforestation in the Amazon rainforest using advanced deep learning techniques. The project leverages neural networks to analyze satellite imagery, comparing images of the same geographic location across consecutive years. Through change detection methodologies, the system identifies new areas of forest loss, providing a faster, scalable solution to support conservation efforts.
The development of automated deforestation monitoring systems is critical for enabling timely interventions by conservationists and policymakers, helping to target high-priority areas effectively. Additionally, this research explores the application of Fourier transform techniques to enhance model accuracy, potentially setting a new standard for change detection in large-scale, unstructured datasets.
This project exemplifies my commitment to applying machine learning and AI to address urgent global challenges. By contributing to sustainable forest management and conservation, this work represents a meaningful step toward leveraging technology for environmental preservation.