Research On Correlation Between Street View Measurement And Stay Willingness In Harbin Based On Visual Interacton

Urban construction has officially entered the historical stage of transformation and development in China today.As an important urban public space,the focus of urban street has gradually shifted from the descriptive study of the object "object" to the study of the comfort and well-being of the subject "human".The research on the measurement of street view based on the humanistic perspective is gradually rising.In this paper,street view image data as the spatial carrier,through the res UNET convolution neural network image processing technology,the online evaluation interactive platform of street space parking intention is constructed.The correlation between street view measurement index and parking intention at 3066 test points of 1151 streets in Harbin city is studied,and the corresponding optimization strategies for street landscape update are proposed.The results show that: 1)the visual interactive platform for online evaluation of Harbin Street Based on parking intention has a high accuracy(87%);2)the spatial distribution aggregation trend of street landscape measurement index and parking intention is obvious,among which,the street height width ratio(SSR)value of 0.2-0.8 is the most(89.3%);the street walking space ratio(WSR)value is 0-0.077,the most(80.9%);the vehicle is dry The results showed that the disturbance index(VII)value was 0.1-0.63(89.3%);the sky vision index(SVI)value was 0.14-0.44(74.95%);the green vision index(GVI)value was 0-0.22(86.7%);the parking intention score was the highest between 0.055-0.094(79.97%);3)the Pearson correlation between street view measurement index and street parking intention was weak,and at three levels(0-0.4,0.4-0.8,0.8-1.0)Moran’s I was 0.069,0.020 and0.041,respectively.On the basis of correlation analysis,this paper puts forward the strategies and guidelines of street view optimization based on the interaction of street view measurement and parking intention,which provides important data support for government landscape architects to study the impact of street view physical measurement indicators on parking intention,and to improve the quality of street view more efficiently,quickly and accurately.