基于机器学习的历史街区道路植物景观视觉效果评价方法
DOI:
作者:
作者单位:

东北农业大学园艺园林学院

作者简介:

通讯作者:

中图分类号:

TU986

基金项目:

黑龙江省自然科学(LH2021E006),黑龙江省艺术科学规划项目(2023B112)


A machine learning-based method for evaluating the visual effect of roadway plantscapes in historic districts
Author:
Affiliation:

College of Horticulture and Landscape Architecture, Northeast Agricultural University

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    以往的视觉效果研究主要集中在对城市整体环境的评估,而缺少针对城市内部历史街区的研究。为评价历史街区的植物景观视觉效果,本研究结合街景图像和机器学习方法进行分析。通过比较不同模型的性能,选取了ResNeSt模型应用于植物景观协调程度和健康程度的评价研究中。结果表明,ResNeSt模型在分类和回归任务中性能最佳,其评分结果与专家评分一致,并与公众评分具有中高度相关性。同时,植物景观视觉效果受经济因素影响显著,并且街区外侧道路的视觉效果评分普遍高于内侧道路。可见,机器学习模型在历史街区植物景观视觉效果评价中具有较高的有效性,能够为历史街区植物景观的保护与优化提供科学依据,对城市规划和旅游业具有重要应用价值。

    Abstract:

    Previous studies on visual effects primarily focus on evaluating the overall urban environment, lacking specific research on historical districts within cities. In order to evaluate the visual effects of plantscapes in historic dis-tricts, street view images and machine learning methods were used. The ResNeSt model was selected to assess the coordination and health of plantscapes. The results show that the ResNeSt model performs best in classification and regression tasks. Its scores are consistent with expert evaluations and moderately to highly correlated with public evaluations. Additionally, the visual effects of plantscapes are significantly influenced by economic factors, with the visual effect scores of streets outside the historic districts generally higher than those inside. It is con-cluded that machine learning models are highly effective in evaluating the visual effects of plantscapes in historic districts. This provides a scientific basis for their protection and optimization, with important implications for urban planning and tourism.

    参考文献
    相似文献
    引证文献
引用本文

陈奕多,胡海辉. 基于机器学习的历史街区道路植物景观视觉效果评价方法[J]. 科学技术与工程, , ():

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2024-05-09
  • 最后修改日期:2024-06-08
  • 录用日期:2024-07-09
  • 在线发布日期:
  • 出版日期: