Abstract: In the context of high-quality development of the urban and rural living environment in the new era, measuring the perception of urban streetscape blue and green spaces in both two and three dimensions can help provide a scientific reference basis and optimization strategies for the improvement of street space quality and urban greening construction. Taking Shaoxing, Zhejiang, a water town in the south of the lower reaches of the Yangtze River, as an example, this research identifies blue and green spaces along the street in the area within the Second Ring Road by virtue of streetscape image data and the fully convolutional network (FCN) semantic segmentation method and, in combination with the normalized difference vegetation index and the normalized difference water index based on satellite remote sensing images, compares the differences between two-dimensional and three-dimensional blue and green spaces along the street. The results show that: 1) green spaces along the street in Shaoxing are well perceivable overall, despite large spatial variation between different green spaces; 2) although there exist abundant blue spaces along the street, water cannot be seen from many waterfront streets due to visual obstacles arising from roadside vegetation and fences; 3) the proportions of green and blue spaces along the street are moderately correlated in two dimensions and weakly correlated in three dimensions. The research proposes corresponding suggestions for enhancing and improving blue and green spaces along the street in Shaoxing.