测绘通报 ›› 2026, Vol. 0 ›› Issue (3): 156-161.doi: 10.13474/j.cnki.11-2246.2026.0326

• 技术交流 • 上一篇    下一篇

地理文化空间约束AIGC赋能乡村特色建筑风貌生成路径——以广州市增城区为例

邓毛颖1, 曹凯滨2,3, 刘博华2,3, 娄玉婷3,4, 赵元2,3   

  1. 1. 广州市规划和自然资源局, 广东 广州 510000;
    2. 城乡院(广州)有限公司, 广东 广州 510000;
    3. 广东省城乡规划建设智能服务工程技术研究中心, 广东 广州 510000;
    4. 广州市图鉴城市规划勘测设计有限公司, 广东 广州 510000
  • 收稿日期:2025-08-18 发布日期:2026-04-08
  • 通讯作者: 赵元。E-mail:giszy@163.com
  • 作者简介:邓毛颖(1973—),男,博士,教授级高级工程师,研究方向为城乡发展与规划、交通规划、经济地理。E-mail:dengmy19990831@163.com
  • 基金资助:
    广州市增城区科技计划(2024ZCKJ19)

AIGC-empowered generation of rural architectural characteristics under geocultural spatial constraints: a case study of Zengcheng district in Guangzhou

DENG Maoying1, CAO Kaibin2,3, LIU Bohua2,3, LOU Yuting3,4, ZHAO Yuan2,3   

  1. 1. Guangzhou Municipal Planning and Natural Resources Bureau, Guangzhou 510000, China;
    2. Urban and Rural Institute (Guangzhou)Co., Ltd., Guangzhou 510000, China;
    3. Guangdong Engineering Technology Research Center of Intelligent Service of Urban and Rural Planning and Construction, Guangzhou 510000, China;
    4. Guangzhou Tujian Urban Planning, Survey and Design Co., Ltd., Guangzhou 510000, China
  • Received:2025-08-18 Published:2026-04-08

摘要: 随着乡村振兴战略的纵深推进,采用人工智能生成内容(AIGC)精准适配乡村建设动态化、差异化需求,为保存乡村集体记忆与文化基因、塑造宜居环境、提升乡村吸引力提供了核心支撑。本文以广州市增城区为实证对象,提出了“特征认知—特征学习—特征生成”的技术闭环框架,系统性地开展了地理文化空间基因解码、乡土建筑多模态数据集构建及AIGC生成模型的多尺度场景验证,从而探索AIGC技术驱动乡村特色风貌精准生成的协同机制。成果不仅实现了文化地理约束下的风貌生成可控性,更形成了“数据—算法—应用”三级推进范式,为人工智能时代的乡村文化传承与空间治理提供了技术方法支撑。

关键词: 乡村振兴, AIGC, 地理文化空间, 乡村风貌, 生成模型, 人工智能

Abstract: With the deepening advancement of rural revitalization strategies,AIGC (artificial intelligence generated content),with its rapid response capabilities and diverse content generation,precisely meets the dynamic and differentiated needs of rural construction.It provides core support for preserving rural collective memory and cultural genes,shaping livable environments,and enhancing rural attractiveness.This study,taking Zengcheng district in Guangzhou as an empirical case,proposes a closed-loop technical framework of “feature cognition-feature learning-feature generation.”It systematically implements geo-cultural spatial gene decoding, construction of a multimodal dataset for vernacular architecture,and multi-scale scenario validation of AIGC generative models.The research explores synergistic mechanisms for AIGC-driven targeted generation of distinctive rural landscapes.Key outcomes include:①Achieved controllable landscape generation under cultural-geographical constraints; ②Established a three-tier advancement paradigm (data-algorithm-application); ③ Provided technical methodological support for rural cultural preservation and spatial governance in the AI era.

Key words: rural revitalization, AIGC, geocultural space, rural landscape, generative model, artificial intelligence

中图分类号: