[论文] MonoIR-RS: Infrared Remote Sensing Vision-Language Learning with CLIP ...

论文概要

研究领域: CV 作者: Jiaju Han, Ma Yaqi, Yahui Chai 发布时间: 2025-07-09 arXiv: 2507.06827

中文摘要

红外遥感图像捕获强度结构、物体-背景对比度和光照不变线索，这些在RGB图像中通常不可见。然而，大多数遥感视觉-语言资源和模型关注可见光波段语义，使红外视觉-语言理解未被充分探索。我们引入MonoIR-RS，一个大规模红外遥感视觉-语言数据集和基准，将红外感知数据构建与CLIP风格对比适配和VLM指令微调相结合。基于与FusionRS相同的源池和分割，MonoIR-RS保留红外图像作为模型面向的模态，产生600,000张合成红外图像和59,032条保留的红外感知的标题记录。模型实验使用这种保留的语言监督子集，其标题围绕灰度结构和红外风格对比而非RGB外观重写监督。我们表明，合成红外图像在AVIID基准上比灰度转换明显更接近真实热成像。我们微调了五个CLIP骨干和六个VLM骨干，并将其与零样本行为进行校准：红外感知适配将CLIP平均召回率提升多达12.8点，并将VLM标题的红外线索覆盖率驱动至100%，同时将残余RGB颜色泄漏降至接近零。通过将红外模态从RGB-IR双模态学习中隔离出来，MonoIR-RS为将红外遥感证据与语言对齐提供了一个受控、可复现的测试平台。

原文摘要

Infrared remote-sensing imagery captures intensity structure, object-background contrast, and illumination-invariant cues often invisible in RGB imagery. Yet, most remote-sensing vision-language resources and models focus on visible-band semantics, leaving infrared vision-language understanding underexplored. We introduce MonoIR-RS, a large-scale infrared remote-sensing vision-language dataset and benchmark that couples IR-aware data construction with CLIP-style contrastive adaptation and VLM instruction tuning. Built from the same source pool and split as FusionRS, MonoIR-RS retains the infrared image as the model-facing modality, yielding 600,000 synthesized infrared images and 59,032 retained IR-aware caption records. The model experiments use this retained language-supervision subset, ...

--- *自动采集于 2026-07-09*

#论文 #arXiv #CV #小凯