摘要: |
在中国快速城镇化阶段“重量”而“轻质”的建设过程中,城市建成区尤其是老城区的景观环境产生了诸如景观视廊受阻、风貌破败等大量问题。对此,自2015年中央城市工作会议以来,各地相继开展了城市修补专项规划,以修复及更新城市建成环境,促进城市空间品质提升。然而,由于缺乏对城市环境整体效应的统筹考虑,故在城市实际修补过程中,城市局部地块的品质提升反而对城市整体品质构成负面影响。因此,在人工智能及大数据技术深度介入城市规划和设计实践探索的基础上,研究基于全卷积神经网络模型(FCN)和城市场景要素深度学习数据集,对城市景观环境中的各要素进行了大规模且高颗粒度的精确识别,同时与空间数据叠加,对复杂建成环境中的景观问题进行精确分析,并基于分析成果辅助后续城市规划设计实践,逐层递进地对城市复杂建成环境进行精细化修补。选择位于嵩山脚下的登封市作为案例,探索人工智能技术在辅助城市修补等规划领域的前瞻性应用。 |
关键词: 风景园林 数字化 城市修补 街景大数据 深度学习 精细化 |
DOI: |
|
基金项目: |
|
Research on Elaborate Urban Repair Planning Approach Based on AI Analysis of Large-scale Street-view Big Data |
ZHENG Yi,YANG Junyan |
Abstract: |
During the rapid urbanization process, the urban construction usually focusses on the quantity and ignore the quality of urban space. The landscape environment of urban built-up areas, especially the old urban areas, has caused a lot of problems, such as blocked view corridors and dilapidated landscape. In this regard, since the Central Urban Work Conference in 2015, various regions have launched special urban repair planning to repair and update the built urban environment and promote the regeneration of urban space. However, during the urban repair process, lack of integrated consideration, systematic and accurate evaluation of the current urban environment results overall urban environment is inharmonious. Based on this situation, due to the deep involvement of artificial intelligence and big data technology in urban planning and design practice, 125-layer FCN full convolutional neural network model and the deep learning data set of urban scene elements were adopted to accurately identified the elements in the urban landscape environment on a large scale and with high granularity. Simultaneously, through superimposed the identification results with urban spatial data, the landscape problems in the complex-built environment are precisely analyzed, and the results are applied to the design and planning parts, then repaired planning for urban complex-built environment will be conducted step by step. Furthermore, Dengfeng was selected as a case study to further explore and improve the prospective application of artificial intelligence technology in urban repair and other planning fields. |
Key words: landscape architecture digitization urban repair street images big data deep learning elaborate |