Leveraging Remote Sensing, Object Detection, and Vision-Language Models for Intelligent Disaster Management and Response

Applications are invited for a self-funded, 3 year full-time or 6-year part-time PhD project.

The PhD will be based in the school of computing and will be supervised by Dr Gelayol Golcarenarenji

The work on this project will:

• Develop a pipeline to detect and classify disaster-related objects (e.g., damaged infrastructure, flood zones, trapped humans) from satellite and drone imagery.
• Employ VLMs for automatic and semantic interpretation of detected scenes.
• Generate actionable insights (e.g., response priorities, damage severity summaries) for first responders and relief organisations.

Project description

Disasters triggered by natural forces or human actions pose significant threats to societies worldwide, often leading to cross-border destruction and humanitarian crises. Each year, floods, earthquakes, storms, and explosions result in the loss of countless lives and inflict severe damage on infrastructure and economies. To support disaster response and recovery efforts, remote sensing (RS) technology, which enables large-scale Earth observation from afar, has seen widespread application. It plays a crucial role in assessing and evaluating damages. In light of the urgent nature of disaster management, integrating artificial intelligence into these processes is essential to improve the speed and accuracy of decision-making.

Object detection techniques applied to remote sensing imagery allow automated identification and localisation of critical features such as damaged buildings, flooded areas, and blocked roads, enabling rapid damage assessment. Furthermore, vision-language models (VLMs) provide a powerful tool to enhance situational awareness and support effective communication among responders and decision-makers. Together, these AI-driven approaches promise to significantly enhance disaster management workflows by combining precise visual analysis with contextual understanding. Hence, the aim of this project is to develop an integrated framework that leverages remote sensing imagery, object detection, and vision-language models to enhance disaster damage assessment and response.

General admissions criteria

The entry requirements for a PhD or MPhil include an upper second-class honors degree or equivalent in a relevant subject or a master's degree in an appropriate subject. Exceptionally, equivalent professional experience and/or qualifications will be considered. All applicants are subject to interview.

If English is not your first language, you'll need English language proficiency at a minimum of IELTS band 6.5 with no component score below 6.0.

If you don't meet the English language requirements yet, you can achieve the level you need by successfully completing a pre-sessional English programmed before you start your course.

Specific candidate requirements

You should have computer programming knowledge using Python, Pytorch or Tensorflow.