Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.18493 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915756186271744 |
|---|---|
| author | Tehrani, Sara Xu, Yonghao Haglund, Leif Berg, Amanda Felsberg, Michael |
| author_facet | Tehrani, Sara Xu, Yonghao Haglund, Leif Berg, Amanda Felsberg, Michael |
| contents | Timely interpretation of satellite imagery is critical for disaster response, yet existing vision-language benchmarks for remote sensing largely focus on coarse labels and image-level recognition, overlooking the functional understanding and instruction robustness required in real humanitarian workflows. We introduce DisasterInsight, a multimodal benchmark designed to evaluate vision-language models (VLMs) on realistic disaster analysis tasks. DisasterInsight restructures the xBD dataset into approximately 112K building-centered instances and supports instruction-diverse evaluation across multiple tasks, including building-function classification, damage-level and disaster-type classification, counting, and structured report generation aligned with humanitarian assessment guidelines.
To establish domain-adapted baselines, we propose DI-Chat, obtained by fine-tuning existing VLM backbones on disaster-specific instruction data using parameter-efficient Low-Rank Adaptation (LoRA). Extensive experiments on state-of-the-art generic and remote-sensing VLMs reveal substantial performance gaps across tasks, particularly in damage understanding and structured report generation. DI-Chat achieves significant improvements on damage-level and disaster-type classification as well as report generation quality, while building-function classification remains challenging for all evaluated models. DisasterInsight provides a unified benchmark for studying grounded multimodal reasoning in disaster imagery. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_18493 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DisasterInsight: A Multimodal Benchmark for Function-Aware and Grounded Disaster Assessment Tehrani, Sara Xu, Yonghao Haglund, Leif Berg, Amanda Felsberg, Michael Computer Vision and Pattern Recognition Timely interpretation of satellite imagery is critical for disaster response, yet existing vision-language benchmarks for remote sensing largely focus on coarse labels and image-level recognition, overlooking the functional understanding and instruction robustness required in real humanitarian workflows. We introduce DisasterInsight, a multimodal benchmark designed to evaluate vision-language models (VLMs) on realistic disaster analysis tasks. DisasterInsight restructures the xBD dataset into approximately 112K building-centered instances and supports instruction-diverse evaluation across multiple tasks, including building-function classification, damage-level and disaster-type classification, counting, and structured report generation aligned with humanitarian assessment guidelines. To establish domain-adapted baselines, we propose DI-Chat, obtained by fine-tuning existing VLM backbones on disaster-specific instruction data using parameter-efficient Low-Rank Adaptation (LoRA). Extensive experiments on state-of-the-art generic and remote-sensing VLMs reveal substantial performance gaps across tasks, particularly in damage understanding and structured report generation. DI-Chat achieves significant improvements on damage-level and disaster-type classification as well as report generation quality, while building-function classification remains challenging for all evaluated models. DisasterInsight provides a unified benchmark for studying grounded multimodal reasoning in disaster imagery. |
| title | DisasterInsight: A Multimodal Benchmark for Function-Aware and Grounded Disaster Assessment |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.18493 |