We are pleased to announce that the following abstracts have been accepted for the Spatial AI Challenge 2024, hosted on the I-GUIDE Platform. These submissions exemplify innovation in AI-ready spatial data, machine learning models, and applications, while promoting FAIR data principles and open science practices.

Accepted Abstracts

1. Extraction of Surface Water Features from High-Resolution Orthoimagery and Elevation Data in Alaskan Watersheds using Novel Computer Vision Techniques

• Team Name: Individual Submission
• Team Members:
  • Uku-Kaspar Uustalu, Tufts University

Summary: This project aims to utilize advanced computer vision techniques, including attention mechanisms, transformer-based models, and graph neural networks, to predict hydrographic features such as flowlines and surface water boundaries from elevation data and high-resolution orthoimagery. The approach involves modifying existing segmentation models like SAM (Segment Anything Model) and YOLO to accommodate surface water extraction, potentially achieving high accuracy with reduced computational resources. The project emphasizes reproducible science by documenting the entire workflow in an executable notebook.

2. Scalable Roof Material Classification to Inform Disaster Resilience Using OpenStreetMap Data

• Team Name: Roof Mappers
• Team Members:
  • Julian Huang, University of Chicago
  • Alex Monito Nhancololo, University of São Paulo
  • Yue Lin, University of Chicago

Summary: This project proposes a scalable pipeline for classifying roof materials in Nepal by leveraging OpenStreetMap (OSM) data. Utilizing existing OSM roof material labels and building outlines, the team plans to train a convolutional neural network (CNN) to classify roof materials into categories such as concrete, reinforced cement concrete (RCC), and tin/metal. The model will be applied to areas lacking roof material labels to produce high-resolution, up-to-date roof material maps. The resulting data can inform disaster mitigation strategies and serve as a proxy indicator of housing quality, household income, and climate vulnerability.

3. Creation of a Machine-Learning-Ready High-Resolution RGB+NIR Orthoimagery Dataset with LiDAR-Derived "Ground Truth" Reference Products

• Team Name: Individual Submission
• Team Members:
  • Uku-Kaspar Uustalu, Tufts University

Summary: This project aims to create a machine-learning-ready dataset by combining high-resolution RGB+NIR (red, green, blue, and near-infrared) orthoimage tiles with corresponding LiDAR-derived reference data, such as digital surface models (DSMs), digital terrain models (DTMs), and canopy height models (CHMs). The dataset will be packaged into a Python library for seamless integration with common deep learning and computer vision frameworks, lowering the barrier for geospatial researchers and AI practitioners interested in aerial imagery analysis.

4. Identify Critical Transportation Infrastructure at Risk during Wildfires

• Team Name: Team CNA
• Team Members:
  • Steven Habicht, CNA
  • Matthew Prebble, CNA
  • Lars Hanson, CNA
  • Rebekah Yang, CNA
  • Shaun Williams, CNA
  • Jeremiah Huggins, CNA

Summary: This study focuses on developing a dataset to identify and assess risks to road and rail transportation infrastructure in California regions susceptible to wildfires. By integrating soil moisture analyses, infrared and satellite imagery, and authoritative datasets like HIFLD, the team will use segmentation models, including U-Net, to detect and classify transportation features and track evolving wildfire risk. The goal is to produce an AI-ready dataset and map visualizing transportation infrastructure at risk from wildfire disruption, providing a reproducible process documented in Jupyter Notebooks.

5. A Vision System for Geospatial Data Extraction

• Team Name: Purdue RCAC
• Team Members:
  • Jungha Woo, Purdue Research Computing
  • Elham Jebalbarezi Sarbijan, Purdue Research Computing

Summary: This project aims to develop a vision system capable of understanding map images, extracting bounding box coordinates, and generating captions that describe map content. By fine-tuning models like CLIP and Vision Transformers, the system will interpret legends, scales, symbols, and coordinates in geospatial maps, facilitating automated extraction of spatial information. This advancement will enhance the capabilities of large language models in geospatial reasoning, aiding researchers in efficiently querying and utilizing legacy datasets.

6. A POMDP-Based Multi-Agent Collaborative Framework for Disaster Sensing and Recovery: Applications of Geographic Intelligence

• Team Name: Rayford
• Team Members:
  • Yifan Yang, Texas A&M University

Summary: This study proposes an automated disaster sensing, recovery, and coordination framework based on Partially Observable Markov Decision Processes (POMDP) and multi-agent collaboration. The framework includes a perception agent for extracting and analyzing street-view imagery, a decision-making agent leveraging a POMDP model to optimize repair tasks, and a coordination agent that integrates geographic information to allocate tasks and optimize resources. The system aims to enhance disaster response efficiency in complex urban environments by utilizing both street-view and remote sensing imagery.

7. Enhancing Disaster Perception Through Semantic Modeling: Leveraging Large Language Models and Knowledge Graphs for Multi-Modal Data Integration

• Team Name: Rayford
• Team Members:
  • Yifan Yang, Texas A&M University

Summary: This project proposes a framework that integrates multi-modal data, including remote sensing imagery, street-view images, and user-generated content, to enhance disaster perception. Leveraging large language models (LLMs) and knowledge graph technologies, the system performs semantic modeling to extract actionable insights from unstructured data sources. By combining LLMs' capabilities with the structural clarity of knowledge graphs, the framework aims to improve disaster impact assessments, severity evaluations, and support decision-making in disaster management.

8. TorchSpatial: A Location Encoding Framework and Benchmark for Spatial Representation Learning

• Team Name: Spatially Explicit Artificial Intelligence Lab
• Team Members:
  • Gengchen Mai, University of Texas at Austin
  • Nemin Wu, University of Georgia
  • Qian Cao, University of Georgia
  • Zhangyu Wang, University of California Santa Barbara
  • Zeping Liu, University of Texas at Austin
  • Yanlin Qi, University of California Davis
  • Jielu Zhang, University of Georgia
  • Ni Lao, Google LLC

Summary: TorchSpatial is a comprehensive framework and benchmark suite designed to advance spatial representation learning (SRL). It supports the development and evaluation of location encoders using extensive benchmarks and innovative evaluation metrics. The framework integrates 15 recognized location encoders to enhance scalability and reproducibility. It includes the LocBench benchmark, which comprises 17 datasets for geo-aware image classification and regression, enabling thorough performance assessments across various geographic distributions. Additionally, TorchSpatial introduces the Geo-Bias Score, a novel metric to evaluate model performance and geographic bias, promoting spatial fairness in GeoAI applications.

We congratulate all the authors on their successful submissions and look forward to their contributions to advancing spatial AI and open science. Participants are reminded of the upcoming deadlines:

• Submission Deadline: March 31, 2025
• Open Competition Deadline: April 15, 2025
• Announcement of Winners: May 15, 2025

For more information and updates, please visit the Spatial AI Challenge 2024 page.