| 摘要: |
| 解析中国园林“步移景异”产生形成的感知机制原
理，利用网络大数据识别由视点(观赏的立足点)感知而来的
“景”及与“景”相应之“异”，对视点及其视觉空间界面予
以要素指标量化分析。以苏州古典园林网师园为例，借助网络
大数据、调查问卷精准高效获取所需要素指标数据，利用人工
智能机器学习生成决策树，构建模拟步移景异感知生成的模
型，得到了4条有利步移景异生成的规则和3条不利步移景异
生成的规则。初试结果显示，影响网师园步移景异感知生成指
标的贡献程度大小依次为视野中水体占比的变化、景阔变化、
景深变化、两视点之间空间转换的次数等。旨在科学量化解析
苏州古典园林时空感知的视觉引发生成机理，破解“步移景
异”之谜，推进中国古典园林设计的传承弘扬。 |
| 关键词: 风景园林 步移景异 苏州古典园林 景观时空感
知 景观视觉感知 网络大数据 人工智能与机器学习 |
| DOI： |
|
| 基金项目: |
|
| The Spatio-temporal Perception FormationMechanism of the "View Changes with StepMovements" in the Master of the Nets Garden |
| ZHANG Tiantian,LIU Binyi,ZHU Zhe,FENG Maohuan |
| Abstract: |
| This study focuses on the perceptual mechanism
and principle behind the formation of "view changes with step
movements" in Chinese gardens, by using internet big data to identify
the elements and indicators of the "view" and its corresponding "step
movements", and conducting a quantitative analysis of the factors and
indicators of viewpoints and their visual spatial interface. Taking the
Master of the Nets Garden in Suzhou as an example, the data for the
needed factors and indicators is obtained through big data analysis
and survey questionnaires precisely and efficiently. The artificial
intelligence machine learning is utilized to generate a decision tree
and construct a perception generation model for "view changes with
step movements", resulting in four rules that are conducive and
three rules that are unfavorable to generate such perception. The
research results indicate that the contribution of different indicators
to the perception includes the variation of water bodies' proportion
in the field of view, changes in the width of the view, the variation of
visual depth, and the number of spatial transformations between two
viewpoints. This study aims to scientifically quantify and analyze
the spatio-temporal perception generation mechanisms triggered by
vision in Suzhou classical gardens, to unravel the mystery behind
"view changes with step movements", and promote the inheritance of
classical Chinese garden design. |
| Key words: landscape architecture view changes with step
movements Suzhou classical garden landscape spatio-temporal
perception visual perception of landscape architecture network big
data artificial intelligence and machine learning |
