Assessing nature exposure beyond satellite indices: The impacts of pedestrian space and street vegetation on respiratory health

Abstract Background While urban greening is universally championed for health promotion, its impact on respiratory health in dense contexts presents a complex spatial compromise. This study investigates the spatial heterogeneity of respiratory diseases and evaluates the independent localized effects of streetscapes on health outcomes, aiming to guide precise urban public health interventions. Methods We conducted a spatial epidemiological analysis across 329 communities in central Zhengzhou, China. Moving beyond traditional vegetation indices derived from satellites, we quantified distal environmental exposures from a pedestrian perspective using Street View Imagery processed via a Mask2Former deep learning semantic segmentation framework. Following the identification of disease clusters through spatial autocorrelation, Linear Mixed Models incorporating community level random effects were executed to disentangle the localized pathways of streetscape features on respiratory outpatient visits. Results Spatial analysis revealed a pronounced spatial clustering pattern from core to periphery, with hotspots of high incidence densely localized within aging and dense urban districts. The quantitative modeling identified pedestrian space visibility as the most potent protective factor; a full unit increase was associated with a substantial reduction of 112.01 in disease kernel density (p < 0.05), validating the active protection provided by walkable environments. Conversely, the Green View Index emerged as a significant factor driving risk, where a unit increase corresponded to an elevation of 17.26 in disease kernel density (p < 0.05). This distinct greenery paradox suggests that dense street vegetation in older contexts may trap botanical aeroallergens (e.g., pollen and catkins) and restrict ventilation, directly aggravating respiratory inflammation. Conclusions By linking modifiable local exposures to actual health outcomes, this research transitions from descriptive correlation to mechanistic insights. To mitigate neighborhood health disparities, municipal planning must shift from blindly expanding green quantity to executing targeted ecological solutions, such as selecting hypoallergenic species and pruning canopies. Furthermore, these streetscape metrics should be integrated into local public health surveillance programs to proactively screen vulnerable communities and facilitate localized primary prevention.