<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title>时序建模 - Tag - 堂堂一跑堂</title><link>https://spacetop.win/tags/%E6%97%B6%E5%BA%8F%E5%BB%BA%E6%A8%A1/</link><description>时序建模 - Tag - 堂堂一跑堂</description><generator>Hugo -- gohugo.io</generator><language>zh-CN</language><managingEditor>kingcopper@whu.edu.cn (WangTong)</managingEditor><webMaster>kingcopper@whu.edu.cn (WangTong)</webMaster><lastBuildDate>Mon, 01 Jun 2026 12:00:00 +0800</lastBuildDate><atom:link href="https://spacetop.win/tags/%E6%97%B6%E5%BA%8F%E5%BB%BA%E6%A8%A1/" rel="self" type="application/rss+xml"/><item><title>RSCaMa：首次将Mamba引入遥感变化描述任务，实现高效时空建模</title><link>https://spacetop.win/2026/06/20260601_211500_rscama_change_captioning/</link><pubDate>Mon, 01 Jun 2026 12:00:00 +0800</pubDate><author><name>WangTong</name></author><guid>https://spacetop.win/2026/06/20260601_211500_rscama_change_captioning/</guid><description><![CDATA[<h1 id="rscama首次将mamba引入遥感变化描述任务实现高效时空建模" class="headerLink">
    <a href="#rscama%e9%a6%96%e6%ac%a1%e5%b0%86mamba%e5%bc%95%e5%85%a5%e9%81%a5%e6%84%9f%e5%8f%98%e5%8c%96%e6%8f%8f%e8%bf%b0%e4%bb%bb%e5%8a%a1%e5%ae%9e%e7%8e%b0%e9%ab%98%e6%95%88%e6%97%b6%e7%a9%ba%e5%bb%ba%e6%a8%a1" class="header-mark"></a>RSCaMa：首次将Mamba引入遥感变化描述任务，实现高效时空建模</h1><blockquote>
  <p><strong>论文解读</strong> | IEEE GRSL 2024 | ESI高被引论文</p>
</blockquote><h2 id="-论文信息" class="headerLink">
    <a href="#-%e8%ae%ba%e6%96%87%e4%bf%a1%e6%81%af" class="header-mark"></a>📄 论文信息</h2><table>
  <thead>
      <tr>
          <th>项目</th>
          <th>内容</th>
      </tr>
  </thead>
  <tbody>
      <tr>
          <td><strong>标题</strong></td>
          <td>RSCaMa: Remote Sensing Image Change Captioning with State Space Model</td>
      </tr>
      <tr>
          <td><strong>作者</strong></td>
          <td>Chen-Yang Liu et al.</td>
      </tr>
      <tr>
          <td><strong>会议</strong></td>
          <td>IEEE Geoscience and Remote Sensing Letters (GRSL) 2024</td>
      </tr>
      <tr>
          <td><strong>arXiv</strong></td>
          <td><a href="https://arxiv.org/abs/2405.13366" target="_blank" rel="noopener noreferrer">https://arxiv.org/abs/2405.13366</a></td>
      </tr>
      <tr>
          <td><strong>GitHub</strong></td>
          <td><a href="https://github.com/Chen-Yang-Liu/RSCaMa" target="_blank" rel="noopener noreferrer">https://github.com/Chen-Yang-Liu/RSCaMa</a></td>
      </tr>
      <tr>
          <td><strong>关键词</strong></td>
          <td>遥感变化描述、状态空间模型、Mamba、时序建模、多时相遥感</td>
      </tr>
  </tbody>
</table>
<h2 id="-解决的核心问题" class="headerLink">
    <a href="#-%e8%a7%a3%e5%86%b3%e7%9a%84%e6%a0%b8%e5%bf%83%e9%97%ae%e9%a2%98" class="header-mark"></a>🎯 解决的核心问题</h2><h3 id="问题背景" class="headerLink">
    <a href="#%e9%97%ae%e9%a2%98%e8%83%8c%e6%99%af" class="header-mark"></a>问题背景</h3><p>遥感图像变化描述（Remote Sensing Image Change Captioning, RSICC）是一项新兴的多模态任务，旨在<strong>用自然语言描述多时相遥感图像之间的地表变化</strong>。与传统的二元变化检测（仅判断&quot;变/不变&quot;）不同，RSICC需要输出更丰富的语义信息：</p>
<ul>
<li><strong>变化对象</strong>：建筑物、道路、植被等</li>
<li><strong>变化位置</strong>：在哪里发生了变化</li>
<li><strong>变化动态</strong>：是新增还是消失</li>
</ul>
<h3 id="现有方法的局限" class="headerLink">
    <a href="#%e7%8e%b0%e6%9c%89%e6%96%b9%e6%b3%95%e7%9a%84%e5%b1%80%e9%99%90" class="header-mark"></a>现有方法的局限</h3><ol>
<li><strong>CNN-based方法</strong>：感受野有限，难以捕获长距离时空依赖关系</li>
<li><strong>Transformer-based方法</strong>：自注意力机制的二次复杂度导致计算成本高昂，特别是在处理高分辨率遥感图像时</li>
<li><strong>时序建模不足</strong>：现有方法多采用简单的双分支结构，缺乏对时序信息的深度交互</li>
</ol>
<h3 id="核心问题提炼" class="headerLink">
    <a href="#%e6%a0%b8%e5%bf%83%e9%97%ae%e9%a2%98%e6%8f%90%e7%82%bc" class="header-mark"></a>核心问题提炼</h3><blockquote>
  <p><strong>如何在保持线性计算复杂度的同时，实现双时相遥感图像之间的深度时空交互，从而生成更准确的变化描述？</strong></p>
</blockquote><h2 id="-解决方案" class="headerLink">
    <a href="#-%e8%a7%a3%e5%86%b3%e6%96%b9%e6%a1%88" class="header-mark"></a>💡 解决方案</h2><h3 id="核心创新点1temporal-traversing-ssm-tt-ssm" class="headerLink">
    <a href="#%e6%a0%b8%e5%bf%83%e5%88%9b%e6%96%b0%e7%82%b91temporal-traversing-ssm-tt-ssm" class="header-mark"></a>核心创新点1：Temporal-Traversing SSM (TT-SSM)</h3><p><strong>设计动机</strong>：
Mamba架构的时间扫描特性与RSICC任务的时序需求存在天然契合。传统SSM采用单向扫描，无法充分利用双时相图像之间的交互信息。</p>
<p><strong>具体实现</strong>：
TT-SSM采用<strong>时间交叉扫描策略</strong>，让两个时相的特征在网络中&quot;交错前行&quot;：</p>
<div class="code-block highlight is-open show-line-numbers  tw-group tw-my-2">
  <div class="
    
    tw-flex 
    tw-flex-row
    tw-flex-1 
    tw-justify-between 
    tw-w-full tw-bg-bgColor-secondary
    ">      
    <button 
      class="
        code-block-button
        tw-mx-2 
        tw-flex
        tw-flex-row
        tw-flex-1"
      aria-hidden="true">
          <div class="group-[.is-open]:tw-rotate-90 tw-transition-[transform] tw-duration-500 tw-ease-in-out print:!tw-hidden tw-w-min tw-h-min tw-my-1 tw-mx-1"><svg class="icon"
    xmlns="http://www.w3.org/2000/svg" viewBox="0 0 320 512"><!-- Font Awesome Free 5.15.4 by @fontawesome - https://fontawesome.com License - https://fontawesome.com/license/free (Icons: CC BY 4.0, Fonts: SIL OFL 1.1, Code: MIT License) --><path d="M285.476 272.971L91.132 467.314c-9.373 9.373-24.569 9.373-33.941 0l-22.667-22.667c-9.357-9.357-9.375-24.522-.04-33.901L188.505 256 34.484 101.255c-9.335-9.379-9.317-24.544.04-33.901l22.667-22.667c9.373-9.373 24.569-9.373 33.941 0L285.475 239.03c9.373 9.372 9.373 24.568.001 33.941z"/></svg></div>
          <p class="tw-select-none !tw-my-1">text</p>]]></description></item><item><title>SkySense：20亿参数多模态遥感基础模型，统一理解地球观测</title><link>https://spacetop.win/2026/06/20260601_120000_skysense_multimodal_foundation_model/</link><pubDate>Mon, 01 Jun 2026 12:00:00 +0800</pubDate><author><name>WangTong</name></author><guid>https://spacetop.win/2026/06/20260601_120000_skysense_multimodal_foundation_model/</guid><description><![CDATA[<h1 id="skysense20亿参数多模态遥感基础模型统一理解地球观测" class="headerLink">
    <a href="#skysense20%e4%ba%bf%e5%8f%82%e6%95%b0%e5%a4%9a%e6%a8%a1%e6%80%81%e9%81%a5%e6%84%9f%e5%9f%ba%e7%a1%80%e6%a8%a1%e5%9e%8b%e7%bb%9f%e4%b8%80%e7%90%86%e8%a7%a3%e5%9c%b0%e7%90%83%e8%a7%82%e6%b5%8b" class="header-mark"></a>SkySense：20亿参数多模态遥感基础模型，统一理解地球观测</h1><blockquote>
  <p><strong>论文解读</strong> | CVPR 2024 | 2026-06-01</p>
</blockquote><h2 id="-论文信息" class="headerLink">
    <a href="#-%e8%ae%ba%e6%96%87%e4%bf%a1%e6%81%af" class="header-mark"></a>📄 论文信息</h2><table>
  <thead>
      <tr>
          <th>项目</th>
          <th>内容</th>
      </tr>
  </thead>
  <tbody>
      <tr>
          <td><strong>标题</strong></td>
          <td>SkySense: A Multi-Modal Remote Sensing Foundation Model Towards Universal Interpretation for Earth Observation Imagery</td>
      </tr>
      <tr>
          <td><strong>作者</strong></td>
          <td>Xin Guo, Jiangwei Lao, Bo Dang, Yingying Zhang, Lei Yu, Lixiang Ru, Liheng Zhong, Ziyuan Huang, Kang Wu, Dingxiang Hu, Huimei He, Jian Wang, Jingdong Chen, Ming Yang, Yongjun Zhang, Yansheng Li</td>
      </tr>
      <tr>
          <td><strong>会议</strong></td>
          <td>CVPR 2024</td>
      </tr>
      <tr>
          <td><strong>arXiv</strong></td>
          <td><a href="https://arxiv.org/abs/2312.10115" target="_blank" rel="noopener noreferrer">https://arxiv.org/abs/2312.10115</a></td>
      </tr>
      <tr>
          <td><strong>GitHub</strong></td>
          <td><a href="https://github.com/Jack-bo1220/SkySense" target="_blank" rel="noopener noreferrer">https://github.com/Jack-bo1220/SkySense</a></td>
      </tr>
      <tr>
          <td><strong>关键词</strong></td>
          <td>遥感基础模型、多模态融合、时序建模、地球观测、通用解译</td>
      </tr>
  </tbody>
</table>
<h2 id="-解决的核心问题" class="headerLink">
    <a href="#-%e8%a7%a3%e5%86%b3%e7%9a%84%e6%a0%b8%e5%bf%83%e9%97%ae%e9%a2%98" class="header-mark"></a>🎯 解决的核心问题</h2><h3 id="问题背景" class="headerLink">
    <a href="#%e9%97%ae%e9%a2%98%e8%83%8c%e6%99%af" class="header-mark"></a>问题背景</h3><p>遥感技术已经渗透到我们生活的方方面面——从城市规划、农业生产到灾害监测、环境保护。然而，传统的遥感影像理解技术存在一个根本性缺陷：<strong>每个任务都需要单独训练一个模型</strong>。比如，要检测建筑物变化，需要一个专门的模型；要识别农作物类型，又需要另一个模型；要监测森林覆盖变化，还需要第三个模型。</p>]]></description></item></channel></rss>