# DisasterVL: Evidence-Based Visual Reasoning Dataset and Model for Disaster Scene Understanding

**Authors:** Ali Diba, Aisha Al-Mohannadi, Yin Yang, Muhammad Imran, Ferda Ofli  
**Published:** 9 Nov 2026  
**Links:** [arXiv](https://arxiv.org/abs/2611.12144) · [PDF](https://arxiv.org/pdf/2611.12144) · [This page](https://omen-bit.github.io/ai-paper-daily/paper/2611.12144.html)

## Key numbers

- **25**: paired with 25 million structured annotations covering scene captions
- **38,254**: The test set contains 38,254 images spanning 75 countries and 28 hazard types

## Abstract

We introduce DisasterVL, a dataset and benchmark for evidence-grounded reasoning about disaster imagery. It contains approximately 1.2 million images from 142 countries, paired with 25 million structured annotations covering scene captions, incident and location assessments, spatial relationships, and visual question answering (VQA). The VQA annotations link answers to visible evidence and include questions about both present and absent cues, multiple answer formats, and difficulty labels. Region-level annotations link caption phrases to bounding boxes and spatial relationships to the regions they connect. The test set contains 38,254 images spanning 75 countries and 28 hazard types. Gemini 3.1 Pro audited the test annotations, and human reviewers validated a stratified subset. On this benchmark, the 9B DisasterVLM model outperforms the evaluated general-purpose baselines on captioning, open-ended VQA, incident reasoning, and region-level grounding. Gains in text overlap are larger than gains in evidence scores. We plan to publicly release the dataset, trained models, and code.

---

Notes written by AI from the paper, numbers copied from the abstract. Check the paper before you rely on them. From AI Papers Daily.