城市生活型街道空间视觉 品质的大规模测度

对街道空间视觉品质评价方法进行优化,采集街景图片并对其进行物理和感知层面的视觉品质评价.首先,通过5个指标描述街道骨架,采用机器学习分割技术对绿视率、围合度和天空开阔度进行三维计算,对贴线率和横截面比进行二维计算.然后,通过问卷调查对街景视觉品质进行主观的满意度评价.最后,通过相关性和回归分析,构建回归模型及拟合公式.结果表明:街道空间视觉品质满意度与绿视率和贴线率呈显著正相关,与围合度呈显著负相关,与天空开阔度和横截面比无明显相关性;3个强相关指标对视觉品质的影响程度从大到小为贴线率>围合度>绿视率.

Large-Scale Measurement of Visual Quality of Space in Urban Living Streets

The visual quality evaluation method of street space was optimized, street view pictures were collected, and their visual quality was evaluated on physical and perceptual levels. 5 indexes were used to describe the street skeleton, the machine learning segmentation technology was used to calculate greenery, enclosure and openness in three-dimensions, the street wall continuity and cross-sectional proportion were calculated in two-dimensions. Through the questionnaire survey, the subjective satisfaction of visual quality of street view was evaluated. Through correlation and regression analysis, the regression model and fitting formula were constructed. The results show that satisfaction of visual quality of street space is significantly positively correlated with the greenery and the street wall continuity, and negatively correlated with enclosure, but is not correlated with the openness and cross-sectional proportion. The influence degree of the three strong correlation indexes on visual quality is: street wall continuity > enclosure > greenery.