机构地区:[1]陕西师范大学地理科学与旅游学院,西安710119 [2]陕西省旅游信息科学重点实验室,西安710119
出 处:《地球信息科学学报》2024年第2期514-529,共16页Journal of Geo-information Science
基 金:国家自然科学基金面上项目(42071169、42271468);陕西省重点研发项目(2019ZDLSF07-04)。
摘 要:街道是城市旅游的重要吸引物,探讨街道景观色彩特征对游客情感感知的影响,对城市街道景观合理规划和布局具有重要的参考价值。本研究以西安市主要建成区为案例地,运用全卷积神经网络(FCN)和随机森林(RF)算法,构建街景图像情感感知数据集,基于街景图像利用机器学习对街道景观的色彩特征进行提取,构建色彩量化指标并将其进行空间可视化;最后,运用随机森林回归算法探讨街道景观色彩特征与游客情感感知之间的关系,并得出最佳色彩特征参数。结果表明:(1)游客情感感知具有明显的空间分布格局,美丽和活泼情感由中心区域向外逐渐增加,安全和富有在主城区外二环以内区域得分较高,无聊在该范围内则较低,压抑情感由中心区域向外逐渐降低,游客在非惯常环境中的情感感知与居民在惯常环境中的情感感知在空间分布上具有一定的同质性;(2)街道景观色彩特征与游客情感感知呈现出复杂的非线性关系。色彩复杂度对美丽和活泼的影响小于色彩协调度,对无聊、压抑、安全、富有的影响大于色彩协调度,当色彩复杂度取值为0.86,色彩协调度取值为0.84时,游客在六个维度可以获得较好的情感感知;(3)一般情况下,街道景观色彩特征越显著,越能够带给游客较好的情感感知。研究在理论上印证了环境色彩越协调,游客体验感越好这一结论;在方法上,丰富了街景大数据和机器学习方法在旅游情感领域的应用。本研究为城市管理者了解游客的街道景观视觉偏好以及优化街道景观设计提供参考。Streets are an important attraction for urban tourism.Exploring the influence of street landscape color characteristics on tourists' emotional perception holds important reference value for the rational planning and layout of urban street landscape.This study takes the built-up area within the third ring road of Xi'an city as a study case,and employs the Full Convolutional Neural Network(FCN) and Random Forest(RF) algorithms to construct an emotional perception dataset of street images.We use the streetscape images as the basis to extract the color features of the streetscape using machine learning algorithms,and color quantifiers are constructed and spatially visualized;The RF regression algorithm is used to explore the relationship between streetscape color characteristics and tourists' emotional perception,and the optimal color characteristic parameters are derived.The results show that:(1) There is a distinct spatial distribution pattern of tourists' emotional perception.The emotions of beauty and liveness gradually increase from the central area outward,and emotions of safety and wealth emotions score higher in the area within the second ring road outside the main city.While boring emotions score lower in this area,and depressing emotions gradually decrease from the central area outward.This suggests that the spatial distribution pattern of emotional perception shares somewhat homogeneity between tourists' emotional perception in non-routine environment and residents' perception in familiar environment;(2) The color characteristics of the streetscape show a complex non-linear relationship with tourists' emotional perception.For example,color complexity has less effect on emotions of beauty and liveness compared to color coordination and has a greater effect on emotions of boredom,depression,safety,and wealth than color coordination.Moreover,when the value of color complexity is 0.86 and the value of color coordination is 0.84,tourists can obtain better emotional perception across six dimensions;(3) Under non
关 键 词:城市街道 街景图像 色彩特征 情感感知 机器学习 GIS 全卷积神经网络 随机森林
分 类 号:TU984.2[建筑科学—城市规划与设计] P208[天文地球—地图制图学与地理信息工程]
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...