| 摘要: |
| 城市总体层面的绿色空间结构优化通常涉及大尺度、
大规模绿地增补，在建成区更新中难以实施。街区是城市更新
基本单元，审视街区绿色空间结构优化符合当前精细化治理下
的规划需求，也有助于推动城市绿地规划思维和模式的转换。
为克服传统调查分析方法的范围、效率和精度局限，应用深度
学习技术对街区尺度绿色空间结构类型展开甄别。通过街区空
间形态识别、绿色空间结构识别、绿色空间结构与空间形态关
联特征分析、优化模式评估4个技术环节，对北京市和广州市中
心城区展开实证研究，以期为城市更新进程中街区绿色空间结
构问题诊断和优化策略制定提供新的工具和解决方案 |
| 关键词: 风景园林 深度学习 绿色空间 结构 适配
性 街区 城市更新 |
| DOI：10.19775/j.cla.2024.10.0062 |
| 投稿时间：2023-06-22修订日期：2023-12-04 |
| 基金项目:国家自然科学基金面上项目(51978146) |
|
| Using Artificial Intelligence Approaches to IdentifyBlock-Scale Green Space Structure and Derive ItsOptimal Mode |
| ZHOU Conghui,LIU Bowen,ZHANG Shining,QIAN Yucen,WANG Yixuan |
| Abstract: |
| The optimization of green space structure in the city level
usually involves large-scale and abundant green space supplementation,
which is difficult to implement in the context of urban renewal. Block is
the basic operating unit of urban renewal. Accordingly, examining the
optimization of block-scale green space structure meets the requirements
of urban refined governance in current urban renewal and helps to
promote the transformation of urban green space planning thinking and
models. To overcome the constraints in traditional analytical methods,
this study utilized deep learning technology to identify the types of
green space structure at the block scale. Through four technical steps:
identifying the types of blocks' spatial morphology, identifying the types
of blocks' green space structure, analyzing the association patterns of
blocks' green space structure and spatial morphology, and assessing the
adaptability of blocks' green space structure and spatial morphology,
empirical research was conducted on the central urban cities of Beijing
and Guangzhou. The research results provide new tools, ideas, and
solutions for diagnosing and optimizing the block-scale green space
structure in the process of urban renewal |
| Key words: landscape architecture deep learning urban green
space structure adaptability block urban renewal |
