首页|基于WallaceiX与Stable Diffusion的

基于WallaceiX与Stable Diffusion的

扫码查看
人工智能正以迅猛的速度逐步渗透到人类生活的各个方面,建筑设计领域亦是如此。文章利用WallaceiX的遗传算法基础结合Stable Diffusion的人工智能渲染能力,以建筑任务书中的各种控规条件和建筑师对场地的理解作为控制参数,在指定场地内进行自动化的枚举式设计,后通过人为筛选,对自动生成的体块进行AI渲染,仅用时 28 s就获得了 6 张建筑效果图。且通过调节控制类关键词,该软件还可以灵活变换设计风格。相较于传统仅依赖脑力和想象完成前期方案设计的设计策略,这种设计方法能够大幅度缩短建筑项目投标前期的方案推敲时间,有效提高方案比选的效率和科学性。在时间紧迫的项目投标工作中,该设计策略具有较强的适用性。
Study on the Strategies for Rapidly Developing Architectural Conceptual Program Based on WallaceiX and Stable Diffusion
The artificial intelligence is rapidly infiltrating various aspects of human life,and the field of architectural design is without exception.Utilizing the genetic algorithm foundation of WallaceiX,and combined with the artificial intelligence rendering capabilities of Stable Diffusion,this paper employs a range of planning conditions outlined in architectural briefs and the architect's understanding on the site as control parameters,and conducts automated enumerative design within a specified site,followed by manual selection and AI rendering of the automatically generated volumes.Remarkably,this process yields six architectural renderings in merely 28 seconds.Moreover,by adjusting control keywords,the software exhibits flexibility in altering design styles.Compared with the traditional design strategies relying solely on intellectual capacity and imagination for the preliminary project design,this approach significantly reduces the time invested in deliberating on architectural proposals during the pre-bid period,and effectively enhances the efficiency and scientific rigor of the proposal selection process.This design strategy proves particularly applicable in time-sensitive project bidding scenarios.

genetic algorithmartificial intelligenceparametric design

王怡平、孙石村

展开 >

北京工业大学城市建设学部

遗传算法 人工智能 参数化

2025

城市建筑
黑龙江科学技术出版社

城市建筑

影响因子:0.761
ISSN:1673-0232
年,卷(期):2025.22(1)