WildFireVQA: A Large-Scale Radiometric Thermal VQA Benchmark for Aerial Wildfire Monitoring

2026-04-22

AI Summary

This work addresses the absence of benchmarks evaluating multimodal reasoning based on thermal radiation in aerial visual question answering (VQA) for wildfire monitoring. The authors introduce the first large-scale aerial wildfire VQA benchmark that integrates RGB imagery with radiometric thermal imaging, encompassing six task categories and featuring paired thermal-RGB data for the first time. Annotation quality is ensured through a hybrid approach combining multimodal large language model generation with sensor-driven deterministic labeling, supplemented by manual verification and intra- and inter-frame consistency checks. Experimental results demonstrate that current models perform best using RGB inputs alone; however, powerful multimodal architectures exhibit significant performance gains when thermal context is incorporated, highlighting the value of temperature-aware reasoning and revealing both its potential and limitations in safety-critical applications.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Wildfire monitoring requires timely, actionable situational awareness from airborne platforms, yet existing aerial visual question answering (VQA) benchmarks do not evaluate wildfire-specific multimodal reasoning grounded in thermal measurements. We introduce WildFireVQA, a large-scale VQA benchmark for aerial wildfire monitoring that integrates RGB imagery with radiometric thermal data. WildFireVQA contains 6,097 RGB-thermal samples, where each sample includes an RGB image, a color-mapped thermal visualization, and a radiometric thermal TIFF, and is paired with 34 questions, yielding a total of 207,298 multiple-choice questions spanning presence and detection, classification, distribution and segmentation, localization and direction, cross-modal reasoning, and flight planning for operational wildfire intelligence. To improve annotation reliability, we combine multimodal large language model (MLLM)-based answer generation with sensor-driven deterministic labeling, manual verification, and intra-frame and inter-frame consistency checks. We further establish a comprehensive evaluation protocol for representative MLLMs under RGB, Thermal, and retrieval-augmented settings using radiometric thermal statistics. Experiments show that across task categories, RGB remains the strongest modality for current models, while retrieved thermal context yields gains for stronger MLLMs, highlighting both the value of temperature-grounded reasoning and the limitations of existing MLLMs in safety-critical wildfire scenarios. The dataset and benchmark code are open-source at https://github.com/mobiiin/WildFire_VQA.

Problem

• wildfire monitoring
• visual question answering
• radiometric thermal data
• aerial imagery
• multimodal reasoning

Innovation

• radiometric thermal imaging
• aerial wildfire monitoring
• multimodal VQA
• multimodal LLM evaluation
• sensor-grounded reasoning