NDVI와 고해상도 식생 마스크를 활용한 계절별 Green View Index 추정 = Estimating Seasonal Green View Index via Fusion of Sentinel-2 NDVI Time Series and High-Resolution Vegetation Mask

다국어 초록 (Multilingual Abstract)

The Green View Index (GVI) quantifies the visibility of greenery from a pedestrian perspective and is widely used to assess walking environments. As the importance of pedestrian-view visual greenery has grown, there is a need to analyze seasonal diversity beyond the greenery-rich summer months. GVI is typically derived from street-view imagery; however, such imagery is often captured intensively in specific seasons and updated at roughly annual intervals, limiting its ability to reflect seasonal variability. This study proposes a framework for estimating the seasonal Green View Index (GVI) by fusing a Sentinel-2 NDVI time series with a high-resolution vegetation mask. Using the proposed approach, which incorporates seasonal NDVI, high-resolution vegetation cover, and cyclical month encoding, we achieved strong performance (R2 = 0.819, MAE = 0.038, r = 0.906).
Improvements were especially pronounced in spring and autumn, when vegetation vigor is low but visually perceived greenery is salient. These results demonstrate that augmenting traditional satellite-based greenness metrics (NDVI) with high-resolution structural vegetation information and acquisition month substantially enhances the estimation of pedestrian-view GVI and its seasonal variability.