| 摘要: |
| 历史街区是地域建筑风貌特征
的重要载体，然而在历史街区风貌保护
与发展过程中，由于不同责任主体对建
筑立面的传统性认知存在偏差，可能会
导致具有重要地域特征的风貌逐渐丢
失，依赖人工智能对现存建筑立面进行
风貌提取，又可能会导致要素识别不清
晰、风貌特征冗余或缺失，从而影响传
统风貌的准确性。研究以苏埠明清老街
风貌整治项目为契机，选择皖南和皖西
典型街区为例，首先系统梳理地域传统
风貌要素，进而通过感性评价与粗糙集
理论探索构成传统特征风貌的要素集合
与组合规则，并尝试将组合规则结合人
工智能与数字化生成技术，运用于苏埠
明清老街风貌更新设计。研究旨在建立
历史街区立面风貌的表述与分析方法，
并联动传统风貌特征的要素提取与数字
化生成设计，为街区的风貌传承与再生
提供借鉴。 |
| 关键词: 历史街区 粗糙集 人工智
能 组合规则 立面更新 |
| DOI：10.13791/j.cnki.hsfwest.20250519003 |
| 分类号: |
| 基金项目:安徽省哲学社会科学规划项目（AHSKQ2024D134） |
|
| Research on analysis of composition rules and digital generation design of historic districtfacades：A case study of the old Ming and Qing Dynasty Street of Subu, Lu’an City |
|
SUN Xia,LI Zao,LIU Yuting,LV Zexi,SHAO Wei
|
| Abstract: |
| Historic districts, as spatial carriers of regional culture, are important representations of traditional
architectural styles. Building facades, in particular, play a critical role in conveying regional identity, historical
continuity, and cultural values. However, in current practices of style preservation and urban renewal, there
exist numerous challenges. On one hand, different stakeholders—such as government officials, design
institutions, and construction teams—often hold divergent understandings of what constitutes “traditionality”.
These perceptual deviations frequently lead to inconsistent interpretations of regional characteristics, resulting
in the simplification, distortion, or even disappearance of traditional styles. On the other hand, with the rapid
advancement of artificial intelligence (AI) and digital technologies, the extraction and generation of
architectural styles have become increasingly reliant on technical tools. While these technologies significantly
improve efficiency and reduce manual errors, over-reliance on algorithms—without fully accounting for the
complexity and diversity of style-related data sources in historic districts—can result in unclear element
recognition, redundant or missing features, and ultimately compromise the accuracy and authenticity of
traditional style expression.In response to these issues, this study takes the facade improvement project of the
Old Ming and Qing Dynasty Street in Subu, Lu’an City, Anhui Province, as a research opportunity. It selects
representative historic districts from both southern and western Anhui—namely Tunxi Old Street, Wan’an Old
Street, the Old Street of Maotanchang, and Subu Old Street—as case samples. Through field investigation and
facade image collection, the study systematically identifies and catalogs the constituent elements of their
architectural styles, with the aim of uncovering the shared compositional features of traditional facades across
different regional contexts. Based on this data, a multidimensional dataset was constructed, including features
such as doors and windows, wall materials, eaves decorations, and shading structures. A perceptual evaluation
was conducted through structured questionnaires to gather public opinions on the perceived “traditionality” of
various facade compositions, which served as decision values for the subsequent analytical phase.The study
introduces rough set theory as a data analysis framework. By applying attribute reduction techniques, the
method extracts the most representative features from complex datasets, identifying the key combinations of
elements that constitute the traditional facade style. A traditionality evaluation decision matrix was established
to correlate perceptual data with specific facade elements. Analysis results revealed that the ground floor doors,
second-story walls, first-story walls, eaves decorations, and overall facade structure were considered the most
influential factors in the perception of traditionality. Among these, walls and doors—due to their prominent
visibility within the human visual field—emerged as the most dominant visual cues. Further comparative
analysis between the southern and western Anhui samples showed that facades in southern Anhui tend to be
more richly decorated and exhibit more complex combinations of elements, whereas those in western Anhuifeature simpler ornamentation and more concise composition rules. These differences reflect the diverse expressions of traditional architecture across different
regional settings.Building upon the identification of key elements and their combination rules, the study further attempts to integrate the results of rough set analysis
with digital generation technologies. By using the extracted traditional facade features as input parameters and controlling the form-language through algorithmic
processes, several updated facade designs for Subu’s Old Street were generated using machine learning models and digital modeling platforms. The generated
outcomes did not only demonstrate a high degree of visual consistency with public perceptions of traditional style but also exhibited strong design feasibility and
adaptability.The research confirms that rough set theory serves as an effective tool for identifying and constructing the rules governing traditional architectural
features. When combined with digital generation technologies, this method enhances both the systematization and efficiency of style design, offering a new pathway
for the preservation and innovation of historic district facades. Compared with conventional approaches that rely heavily on professional experience or subjective
judgment, this integrated method provides more robust data support and operational feasibility. It offers a scientifically grounded framework for the expression,
preservation, and revitalization of historic architectural styles in the context of contemporary urban development.In summary, this study, based on field investigations
and perceptual evaluations, establishes a logical pathway between traditionality perception and facade style generation. It proposes a complete workflow that spans
from the extraction of style elements, to the construction of combination rules, and ultimately to the design generation process. The research offers a referenceable
paradigm for the preservation and regeneration of historic districts, and holds significant theoretical and practical value in promoting the digital expression of
traditional architectural culture and stimulating the vitality of urban renewal. |
| Key words: historic districts rough sets artificial intelligence combination rules facade renewal |
