用生成模型根据烧毁区域图合成真实卫星影像，提升野火检测数据质量。
Generating Satellite Imagery Data for Wildfire Detection through Mask-Conditioned Generative AI
- 用烧毁掩码引导扩散模型生成野火后卫星图像，分全图生成和上下文补全两种方式。
- 基于上下文补全的方案在烧毁区域匹配度上达到0.456的最高指标，颜色差异最小为63.22。
- 借助视觉语言模型自动生成提示词，效果接近人工设计，适合快速部署应用。
标注过的卫星影像稀缺，严重制约基于深度学习的野火监测系统发展。本文探究一种基于扩散的地球观测基础模型EarthSynth，能否在不进行任务特定微调的前提下，仅以现有烧毁掩码为条件，合成真实的灾后Sentinel-2 RGB影像。利用来自CalFireSeg-50数据集（Martin et al., 2025）的烧毁掩码，设计并评估了六种受控实验配置，系统性地改变：(i) 流水线架构（仅掩码生成全图与带灾前上下文的修复补全），(ii) 提示工程策略（三种人工设计提示与通过Qwen2-VL生成的提示），(iii) 区域级色彩匹配后处理步骤。在10个分层测试样本上，采用四项互补指标（烧毁交并比、烧毁区颜色距离ΔC_burn、暗度对比度、光谱合理性）进行定量评估。结果表明，基于修复补全的流水线在所有指标上均优于全图生成；结构化提示补全方案实现最佳空间对齐（Burn IoU = 0.456）与烧毁显著性（Darkness Contrast = 20.44），而色彩匹配使颜色距离最低（ΔC_burn = 63.22），但烧毁显著性下降。基于视觉语言模型的提示生成方法表现可媲美人工设计。这些发现为将生成式数据增强集成至野火检测流程提供了依据。代码与实验详见：https://www.kaggle.com/code/valeriamartinh/genai-all-runned
原文摘要 · Abstract (English)
The scarcity of labeled satellite imagery remains a fundamental bottleneck for deep-learning (DL)-based wildfire monitoring systems. This paper investigates whether a diffusion-based foundation model for Earth Observation (EO), EarthSynth, can synthesize realistic post-wildfire Sentinel-2 RGB imagery conditioned on existing burn masks, without task-specific retraining. Using burn masks derived from the CalFireSeg-50 dataset (Martin et al., 2025), we design and evaluate six controlled experimental configurations that systematically vary: (i) pipeline architecture (mask-only full generation vs. inpainting with pre-fire context), (ii) prompt engineering strategy (three hand-crafted prompts and a VLM-generated prompt via Qwen2-VL), and (iii) a region-wise color-matching post-processing step. Quantitative assessment on 10 stratified test samples uses four complementary metrics: Burn IoU, burn-region color distance (ΔC_burn), Darkness Contrast, and Spectral Plausibility. Results show that inpainting-based pipelines consistently outperform full-tile generation across all metrics, with the structured inpainting prompt achieving the best spatial alignment (Burn IoU = 0.456) and burn saliency (Darkness Contrast = 20.44), while color matching produces the lowest color distance (ΔC_burn = 63.22) at the cost of reduced burn saliency. VLM-assisted inpainting is competitive with hand-crafted prompts. These findings provide a foundation for incorporating generative data augmentation into wildfire detection pipelines. Code and experiments are available at: https://www.kaggle.com/code/valeriamartinh/genai-all-runned
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。