five

CGAN Assists the Styles & Features Renovation of Street Facade—A case study of Wuyi area in Fujian, China(Training set for machine learning)

收藏
Mendeley Data2024-03-27 更新2024-06-26 收录
下载链接:
https://data.mendeley.com/datasets/k9vfmdtbxm
下载链接
链接失效反馈
官方服务:
资源简介:
This dataset is about the paper "CGAN Assists the Styles & Features Renovation of Street Fa-cade—A case study of Wuyi area in Fujian, China".In the course of the research, we used a large number of vector maps of residential renovation in the Wuyishan area of China for training.Our study proposes a CGAN-based approach. The main purpose is to realize the style migration from local architectural style to village and town building facades, and to provide new design, decoration and restoration of building facades for the village and town architectural style. It has important practical value for the unified style and appearance management of large-scale and multi-building villages and towns.Our sample training images are divided into 3 types: building outline (BEP), functional zoning (FSL), and building facade (BE).Divided into 164 pictures, a total of 492 pictures. The BEP sample uses a grayscale image, and the color is black (R0, G0, B0). In the processing of FSL samples, in order to simplify the data, various elements in the building facade are represented by different colors and presented in the form of color pictures. In this study, 5 colors are used to represent the elements in the facade, that is, the door is red (R255, G99, B128), the window is green (R123, G132, B61), the wall is yellow (R243, G219, B181), and the guardrail is Blue (R3, G172, B213), brown roof (R171, G73, B64).

本数据集围绕学术论文《CGAN辅助街道立面风格与特征改造——以中国福建武夷地区为例》构建。研究过程中,我们采用中国武夷山地区大量住宅改造矢量地图开展模型训练。本研究提出一种基于条件生成对抗网络(Conditional Generative Adversarial Network, CGAN)的方法,核心目标是实现地域建筑风格向村镇建筑立面的风格迁移,为村镇建筑立面的创新设计、装饰及修复提供新思路,对大规模多楼栋村镇的风貌统一管控具有重要实用价值。本研究的训练样本图像分为三类:建筑轮廓(Building Outline, BEP)、功能分区(Functional Zoning, FSL)与建筑立面(Building Facade, BE),共计164组样本,总图像量达492张。其中BEP样本采用灰度图像,色彩设置为纯黑(R0, G0, B0)。在FSL样本的处理流程中,为简化数据体量,建筑立面的各类元素通过差异化色彩进行表征,并以彩色图像形式呈现。本研究共使用5种色彩表征立面元素:门为红色(R255, G99, B128)、窗为绿色(R123, G132, B61)、墙体为黄色(R243, G219, B181)、护栏为蓝色(R3, G172, B213),屋顶为棕褐色(R171, G73, B64)。
创建时间:
2024-01-23
二维码
社区交流群
二维码
科研交流群
商业服务