摘要
本文研究纽约曼哈顿中城的建筑色彩特征,在方法上突破了现实局限性,以街景图片数据为原材料,辅以城市用地数据,借用深度学习的图像分类技术,对研究范围内建筑色彩的主色调进行提取和分析。结果表明:(1)利用街景图片分析国外城市建筑色彩现状的关键技术包括街景图片采集、图像要素语义分割和色彩提取;(2)曼哈顿中城建筑色彩总体以纽约砖红和曼哈顿灰为主,此色彩基因主要由建造年代、区域功能和建筑高度等因素决定;(3)城市色彩规划需要充分尊重地区历史发展规律,保护好城市的优秀历史风貌是做好色彩规划的基础。
This paper studies the color characteristics of buildings in midtown Manhattan,New York,and breaks through the practical limitations in methods.Using street view images as the basic data,supplemented by urban land use data,and borrowing the image classification technology of deep learning,this paper extracts and analyzes the main architectural colors within the research scope.The results show:①the key technologies of analyzing the architectural color status of foreign cities with street view pictures include street view images collection,semantic segmentation of image elements and color extraction.②The architectural colors of midtown Manhattan are mainly New York-brick-red and Manhattan-gray.This color gene is mainly determined by factors such as architectural era,regional function and building height.③The urban color planning needs to fully respect the regional historical development law,and to protect the excellent historical features of the city,which is the basis of excellent color planning.
出处
《城市建筑》
2022年第11期142-145,175,共5页
Urbanism and Architecture
关键词
街景图片
商务区
建筑色彩
色彩规划
street view images
business districts
architectural color
color planning
