A Classification of Tidal Flat Wetland Vegetation Combining Phenological Features with Google Earth Engine

The composition and distribution of wetland vegetation is critical for ecosystem diversity sustainable development. However, tidal flat environments are complex, obtaining effective satellite imagery challenging due to the high cloud coverage. Moreover, it difficult acquire phenological feature data extract species-level information by using only spectral or individual images. To solve these limitations, statistical features, temporal features multiple Landsat 8 time-series images obtained via Google Earth Engine (GEE) platform were compared from Chongming Island, China. results indicated that (1) a harmonic model characteristics better than raw index (VI) Savitzky–Golay (SG) smoothing method; (2) classification based on combination three provided highest overall accuracy (85.54%), (represented amplitude phase model) had greatest impact classification; (3) result senescence period was more accurate green period, but annual mapping all seasons most accurate. method described in this study can be applied overcome impacts complex environment wetlands effectively classify species GEE. This could used as reference analysis other areas types.