Research on Community Perception Based on Street View Data Scoring Strategy

People are the main users of community street space. How to quantify street space quality from the perspective of people's perception and explore the relationship between street space quality and street elements has been a topic devoted to research in various fields. The development of big data and computer science provides a new technique for quantitative evaluation of street perception. The downtown area of Haikou city was taken as the research object to verify the method of this study. Street images of this area were used to construct a deep learning scoring model, and six dimensions of beauty, rich, safety, energetic, repression and [1] boredom were used to score perceptions. The six dimensions of perception are further divided into positive perception and negative perception. The top 20% of streetscape with the highest positive perception score was regarded as high-quality street space, while the top 20% of streetscape with the highest negative perception score was regarded as low-quality street space. The results showed that the highly urbanized area (Longhua district) was considered to be safer and more energetic, but at the same time, with higher sense of depression, there were significant differences incommunity perception among different types of POI, and community green space could significantly relieve the perception of depression. Therefore, this paper confirms the key role of urban functions in promoting sustainable urban development from the perspective of subjective perception, and provides scientific support for planning policies in terms of infrastructure service supply. By exploring the correlation between urban perception and urban functional pattern, this study finds that there is a strong nonlinear relationship between urban perception and urban functional structure. The case study quantitatively determines the impact of urban functions on city perception and illustrates the effectiveness of the proposed deep learning scoring method in promoting city perception evaluation.