基于改进U-Net的历史建筑水刷石饰面配比智能分析技术  

Intelligent Analysis of Water-brushed Stone Facing Ratio in Historical Buildings Based on Improved U-Net

作　　者：陈雪峡 沈俊凯 孙沈鹏 张英楠 谷志旺 CHEN Xuexia;SHEN Junkai;SUN Shenpeng;ZHANG Yingnan;GU Zhiwang(Shanghai Construction No.4(Group)Co.,Ltd.,Shanghai 201103,China)

机构地区：[1]上海建工四建集团有限公司,上海201103

出　　处：《建筑施工》2024年第9期1373-1375,1380,共4页Building Construction

基　　金：国家重点研发计划(2023YFC3805505)。

摘　　要：水刷石饰面是上海近现代历史建筑的典型饰面之一。基于人工智能等现代技术,研发基于改进U-Net的水刷石配比智能分析方法,构建水刷石配比智能分析体系,实现典型水刷石饰面材料配比的快速化、智能化获取,解决了典型特征饰面传统材料配比遗失严重,常规的材料配比分析技术面临的试配时间长、过程反复、精确待提升等问题。研究结果可为现代技术赋能历史风貌的保护和传承奠定基础。Water-brushed stone facing is a typical feature of modern historical buildings in Shanghai.Leveraging modern technologies such as artificial intelligence,an intelligent analysis method based on an improved U-Net is developed to analyze the ratio of water-brushed stone facing.This system enables the rapid and intelligent acquisition of typical water-brushed stone material ratios,addressing the significant issues of traditional material ratio loss,extended trial matching times,repetitive processes,and the need for enhanced precision in conventional material ratio analysis techniques.This method also establishes a strong foundation for the preservation and inheritance of historical features empowered by modern technology.

关 键 词：材料配比智能分析 图像数据集 神经网络模型 水刷石饰面 

分 类 号：TU17[建筑科学—建筑理论]