Undergrade Thesis

Enhancing Infrastructure Detection Using Drone Imagery : A Comparative Analysis Using Machine Learning Models with a Custom Bangladesh-Based Dataset for Improved Urban Planning.

Developed and compared advanced machine learning models (YOLOv5, YOLOv7, YOLOv8, and a custom model) to enhance the accuracy of infrastructure detection using drone imagery for urban planning in Bangladesh.

Created a specialized dataset tailored to Bangladesh's urban features, enabling high-precision model training and testing for diverse urban environments and illumination conditions.

Demonstrated the effectiveness of temporal data for monitoring infrastructural changes over time, contributing to more efficient and sustainable urban design predictions and planning strategies.