Residential Window Views from Real-Estate Panoramas: Measurement, Spatial Patterns, and Environmental Associations in Wuhan

Residential window views provide a perspective on urban landscapes that is difficult to observe at large scale. This study uses real-estate panoramas to quantify residential window-view composition, describe its spatial distribution in Wuhan, and examine associations with surrounding resources and building form. We analyzed 12,334 balcony panoramas from 1377 residential coordinates using semantic segmentation and a window view index, supported by measurement validation and sensitivity analyses. Spatial mapping, floor-group comparisons, and spatial regression characterized horizontal patterns, vertical differences, and environmental associations. Window-view composition varied across locations and floor groups, with geographically concentrated water views. Mean tree share was 24.6% in the low-floor group and 8.2% in the high-floor group, while sky share increased and high-rise-building content peaked in the middle-floor group. Tree share was positively associated with neighborhood green cover. Water visibility was positively associated with water cover and proximity and negatively associated with building footprint ratio. Taller surroundings were associated with less sky and more high-rise-building content. Resource associations also varied among floor groups. The workflow supports neighborhood visual assessment and can inform greening and building-layout review at different elevations.