测绘通报 ›› 2026, Vol. 0 ›› Issue (8): 96-102.doi: 10.13474/j.cnki.11-2246.2026.0814

• 学术研究 • 上一篇    

多模态融合框架下的城市功能区识别

张用川1,2, 高洁1,2, 张治清2,3, 周正强4, 王陆潇3, 官冬杰1   

  1. 1. 重庆交通大学智慧城市学院, 重庆 400074;
    2. 自然资源部国土空间规划监测评估预警重点实验室, 重庆 401120;
    3. 重庆市规划和自然资源信息中心, 重庆 401147;
    4. 重庆市铜梁区规划和自然资源信息中心, 重庆 402560
  • 收稿日期:2025-12-02 发布日期:2026-09-12
  • 通讯作者: 高洁。E-mail:622230100006@mails.cqjtu.edu.cn
  • 作者简介:张用川(1986—),男,博士,讲师,主要研究方向为智能城市与空间规划。E-mail:zhangyc@cqjtu.edu.cn
  • 基金资助:
    自然资源部国土空间规划实施监测评价预警重点试验室开放基金(LMEE-KF2024011)

Identification of urban functional zones based on a multi-modal fusion framework

Zhang Yongchuan1,2, Gao Jie1,2, Zhang Zhiqing2,3, Zhou Zhengqiang4, Wang Luxiao3, Guan Dongjie1   

  1. 1. School of Smart City, Chongqing Jiaotong University, Chongqing 400074, China;
    2. Key Laboratory of Monitoring, Evaluation and Early Warning of Territorial Spatial Planning Implementation, Ministry of Natural Resources, Chongqing 401120, China;
    3. Chongqing Municipal and Natural Resources Information Center, Chongqing 401147, China;
    4. Chongqing Tongliang District Planning and Natural Resources Information Center, Chongqing 402560, China
  • Received:2025-12-02 Published:2026-09-12

摘要: [目的] 针对传统城市功能区识别方法在多源异构数据融合能力不足、复杂地形条件下特征表达受限等问题,本文提出了一种基于深度学习的多模态融合框架。[方法] 分别从遥感影像、POI及DEM数据中提取异构特征,设计辅助权重层(AWL)实现辅助数据源重要性的自适应评估;设计双分支注意力模块(DBAM),通过跨模态注意力机制动态平衡主辅数据分支贡献,构建多模态特征融合模块(MFFM),融合深度可分离卷积与Transformer编码器,增强跨模态语义关联建模能力,实现城市功能区的精准识别。[结果] 以重庆市南岸区2019个样本进行验证试验,本文方法总体精度达到88.02%,Kappa系数为0.847。[结论] 本文提出的框架能够有效地提升复杂城市环境下的功能区识别精度。

关键词: 多模态数据, 城市功能区, 特征提取, 深度学习, 城市规划

Abstract: [Purposes] To address the limitations of traditional urban functional zone (UFZ)identification methods,such as insufficient multi-source data fusion and limited feature representation under complex terrain conditions. [Methods] This study proposes a deep learning-based multimodal fusion framework.The model extracts heterogeneous features from Sentinel-2 imagery,POI data,and DEM information.An auxiliary weighting layer (AWL)is designed to adaptively evaluate the contribution of each auxiliary data source.A dual-branch attention module (DBAM)dynamically balances the influence of remote sensing and auxiliary data through a cross-modal attention mechanism.Additionally,a multi-modal feature fusion module (MFFM)is constructed by integrating depthwise separable convolutions with Transformer encoders to enhance cross-modal semantic representation and improve classification of urban areas. [Findings] Experiments conducted in Nan'an district,Chongqing,using 2019 labeled samples across nine UFZ categories demonstrate that the proposed method achieves an overall accuracy of 88.02% and a Kappa coefficient of 0.847. [Conclusions] These results confirm the effectiveness of the framework in UFZ classification for complex urban environments.

Key words: multi-modal data, urban functional zones, feature extraction, deep learning, urban planning

中图分类号: