Object Detection Using Transformers

The people of Malawi have faced numerous natural disasters and climatic shocks in recent years, such as droughts, floods, and landslides. These events, compounded by the impacts of Covid-19 and other global issues, have severely affected the health and well-being of most Malawians. Rural areas, where more than 80% of the population resides, have been particularly hard-hit.

Project Overview

This project focuses on leveraging transformer-based object detection models to address gaps in accurately mapping flood extents and corresponding damages using satellite imagery. Traditional algorithms often miss grass-thatched roofs in rural areas, which are common in Malawi. By using advanced object detection techniques, this project aims to improve the identification of affected populations and buildings.

Key Features

• Transformer-Based Models: Utilizes state-of-the-art transformer architectures for object detection.
• Satellite Imagery Analysis: Focuses on mapping flood extents and damages in rural Malawi.
• Improved Accuracy: Addresses gaps in detecting traditional grass-thatched roofs and other rural structures.

Skills and Technologies

• Computer Vision: Core technology for analyzing satellite imagery.
• Machine Learning: Powers the detection and classification algorithms.
• Object Detection: Focused on identifying buildings and affected populations.
• Deep Learning: Utilizes advanced transformer models for enhanced accuracy.

This project highlights the potential of transformer-based object detection models in addressing real-world challenges, particularly in disaster response and rural development contexts.