摘要
西南少数民族聚落景观具有丰富的文化价值与游憩价值。针对民族聚落景观评价方法中主观性较强、缺少内容、结构定量分析不足等问题,利用传统景观美景度评价法(SBE)与机器学习方法相结合构建回归模型,分析景观要素构成对其视觉质量的影响。结果表明:①聚落景观美景度与图像要素构成存在较强关联,两者之间的量化关系可通过回归方程Y=8.821X1+9.10X2+4.264X3+5.057X4+4.598X5-1.536表示;②从SBE值与图像内容的回归方程中可以看出,侗族聚落景观的美景度与树木占比关联性最强,其次是天空,与道路和田地的关联性最弱。研究从图像的语义元素组合及像素占比角度分析了图像的美景度。通过定量分析模型,可以在较少人工参与的情况下自动识别美景度,有利于拓展景观评价领域的研究思路与技术手段,并为传统侗族聚落规划设计、景观保护与品质提升提供参考。
The landscape of ethnic minority settlements in Southwest China has rich cultural and recreational values.Aiming at the problems of strong subjectivity,lack of content,and insufficient structural quantitative analysis in the evaluation methods of ethnic settlements,the traditional landscape evaluation method(SBE)and machine learning methods are used to construct a regression model to analyze the composition of landscape elements and their visual quality Impact.The results show that:①There is a strong correlation between the scenic beauty of the settlement landscape and the composition of image elements.The quantitative relationship between the two can be determined by the regression equation Y=8.821X1+9.10X2+4.264X3+5.057X4+4.598X5-1.536 means;②From the regression equation of SBE value and image content,it can be seen that the beauty of the Dong settlement landscape has the strongest correlation with the proportion of trees,followed by the sky,and the weakest correlation with roads and fields..This research analyzes the beauty of the image from the perspective of the image's semantic element combination and pixel ratio.Through the quantitative analysis model,the degree of beauty can be automatically identified with less human participation,which is conducive to the expansion of research ideas and technical methods in the field of landscape evaluation,and provides reference for traditional Dong settlement planning and design,landscape protection and quality improvement.
作者
郑文俊
吴曼妮
周茂杰
HENG Wen-jun;WU Man-ni;ZHOU Mao-jie(School of Tourism and Landscape Architecture,Guilin University of Technology,Guilin,Guangxi,541004)
基金
教育部人文社会科学研究青年基金项目(编号:19YJC 850027)
广西自然科学基金面上项目(编号:2018GX NSFAA050068)
广西中青年教师基础能力提升项目(编号:2018KY0250)。
