张弛;蔺倾程;朱炜;李雪峰;肖辉;
ZHANG Chi;LIN Qingcheng;ZHU Wei;LI Xuefeng;XIAO Hui;Department of Control Science and Engineering, Tongji University;Frontiers Science Center for Intelligent Autonomous Systems, Tongji University;

• 1:同济大学控制科学与工程系
• 2:同济大学教育部自主智能无人系统前沿科学中心

摘要(Abstract)：

城市照明是现代城市文明的典型标识，对于交通安全保障、人居环境建设、城市形象表达、经济发展促进等方面发挥重要作用。作为城市照明的重要组成部分，景观照明结合城市历史底蕴与文化特色，能够提升人民幸福感知与艺术审美。本研究将计算机视觉和图像处理方法引入城市照明，构建基于景观照明图像的美学数据集，提取和筛选有效美学特征，优化机器学习算法形成具有准确性、有效性和高精度的粒子群优化算法—反向传播神经网络(PSO-BPNN)景观照明美学评价模型，并针对实际照明工程开展美学效果的客观定量评估。结果显示，平均相对误差(MRE)仅为8.26%,具有客观性和普适性。可见，本研究探索的基于视觉图像的照明美学评价应用方法，可为城市照明发展提供技术支持和参考依据，助力推进数字城市治理现代化的建设进程。
Urban lighting is a typical sign of modern urban civilization and plays a vital role in traffic safety, human living environment construction, urban image expression, and economic development promotion. As a significant part of urban lighting, landscape lighting combines the historical heritage and cultural characteristics of cities, enhancing people's happiness perception and artistic aesthetics. This study introduces computer vision and image processing methods into urban lighting. An aesthetic database based on landscape lighting images is constructed, and the practical aesthetic features are extracted and filtered. Optimized machine learning algorithms form a particle swarm optimization algorithm-back propagation neural network(PSO-BPNN) landscape lighting aesthetic evaluation model with accuracy, validity, and high precision. The objective quantitative evaluation of aesthetic effects is carried out for lighting projects. The results show that the mean relative error(MRE) is only 8.26%, which is objective and generalizable. The visual image based lighting aesthetic evaluation application method explored in this study can provide technical support and reference basis for urban lighting development and help promote the construction process of digital city governance modernization.

关键词(KeyWords)： 景观照明;美学评价;视觉图像;机器学习
landscape lighting;aesthetic evaluation;visual images;machine learning

Abstract:

Keywords:

基金项目(Foundation):

作者(Authors): 张弛;蔺倾程;朱炜;李雪峰;肖辉;
ZHANG Chi;LIN Qingcheng;ZHU Wei;LI Xuefeng;XIAO Hui;Department of Control Science and Engineering, Tongji University;Frontiers Science Center for Intelligent Autonomous Systems, Tongji University;

参考文献(References)：

• [1] 蒋慧勇.绿色照明在城市照明中的应用研究[J].江苏科技信息，2021,38(23):58-60.
• [2] Jin X,Li X,Feng Z,et al.Linking ecological efficiency and the economic agglomeration of China based on the ecological footprint and nighttime light data[J].Ecological Indicators,2020,111:106035.
• [3] 袁樵.璀璨浦江，魅力上海——黄浦江两岸景观照明总体方案设计和实践[J].照明工程学报，2019,30(5):14-15.
• [4] 韩天飞.张家口冬奥村夜景照明方案从规划设计到实现[J].照明工程学报，2022,33(1):219-222.
• [5] 杨波.2022年冬奥会北京赛区场地照明工程应用[J].照明工程学报，2022,33(1):17-24.
• [6] Abry P,Roux S G,Wendt H,et al.Multiscale anisotropic textureanalysis and classification of photographic prints art scholarship meets image processing algorithms[J].IEEE Signal Processing Magazine,2015,32(4):18-27.
• [7] Tan Y,Tang P,Zhou Y,et al.Photograph aesthetical evaluation and classification with deep convolutional neural networks[J].Neurocomputing,2017,228:165-175.
• [8] Zhang J,Miao Y,Yu J.A comprehensive survey on computational aesthetic evaluation of visual art images:metrics and challenges[J].IEEE Access,2021,9:77164-77187.