ABSTRACT
The idea of a 15-minute community living circle emphasizes the importance of the surrounding environment of residential blocks to the elderly. Applying street view images and deep learning technique, the study developed three visual indicators -greenness, openness, and enclosure - which were calculated by the outcomes of semantic segmentation of street view data. Additionally, the study applied the multilevel regression analysis method to examine their relationship with the levels of depression and anxiety among the elderly. The results reveal that the residential communities with good and poor street greening within the 15-minute life circles in Haidian District show mixed distribution. Specifically, areas with high green visual rates are predominantly located near schools, universities, urban parks, and leisure resorts. The residential blocks inside the Fourth-Ring Road had lower values for openness and a higher sense of enclosure compared to areas outside the Fourth-Ring Road. Moreover, the study found a significant positive correlation between an increase in green vision rate within a 15-minute living circle and a decrease in the degree of depression and anxiety in the elderly, while larger values for openness in the area had a negative influence on the mental health outcomes of the elderly.
