An AI-Assisted Mapping Tool for Disaster Management with Humanitarian OpenStreetMap

Background

Disaster management often faces challenges in accurately and efficiently mapping affected areas, especially in remote or inaccessible regions. Humanitarian organizations require advanced tools to improve situational awareness, facilitate better decision-making, and optimize relief efforts.

Objective

• To develop an AI-assisted mapping tool to support humanitarian organizations in mapping disaster-affected areas.
• To enhance the accuracy and efficiency of mapping efforts, especially in regions that are difficult to access.

Approach

The project employed cutting-edge AI methodologies to tackle the problem of mapping disaster-affected areas:

• Techniques Used: Convolutional Neural Networks (CNNs), Transfer Learning, and the RAMP Deep Neural Network.
• Tools: The HOTLib Python library was developed to implement the AI-assisted mapping workflow.
• Data: Aerial imagery was analyzed to predict building footprints and assess infrastructure damage.
• Processes: By leveraging transfer learning and custom neural networks, the team enhanced the AI model’s ability to process data efficiently and with high accuracy.

Results and Impact

• Outcomes:
  • A deep neural network model achieved an impressive 94% average accuracy in predicting building footprints from aerial images.
  • The creation of the HOTLib Python library offered a reference framework for AI-assisted mapping workflows.
• Impact:
  • Humanitarian organizations now have access to highly accurate maps, which improve decision-making during disaster response.
  • The tool significantly enhances the efficiency of mapping operations, ensuring timely availability of critical data for disaster management.
• Final presentation:

Find a detailed technical case study on the project here.

Future Implications

This project paves the way for future innovations in disaster management:

• Policy Influence: The findings could guide the adoption of AI technologies in global humanitarian policies.
• Research Expansion: The AI-assisted mapping tool and methodologies can serve as a foundation for further research, including adapting the model for other natural disaster scenarios or urban planning purposes.