Abstract: The diversity of wetland landscape coverage provides a wide range of habitat selection for birds, which is important for the maintenance of wetland bird diversity. Based on GF₂ images, the habitat types of wetland vegetation were extracted using object oriented classification method. The coverage of wetland vegetation was calculated using dimidiate pixel model. According to the range of coverage fluctuation, five coverage types were divided in the study area. Finally, the statistical analysis on the coverage of different vegetation was performed. The results showed that, compared with the traditional supervised classification method, the overall accuracy of object oriented classification was higher, reaching 91.21%. The wetland vegetation interpreted by remote sensing showed obvious zonal distribution, with the main habitats being reed marsh, Suaeda salsa marsh, and Spartina alterniflora marsh. The distribution pattern of vegetation coverage was high on all sides and low in the center, with differences among various dominant vegetation coverage types. Among them, reed marsh mainly distributed in the superior coverage area and high coverage area, Spartina alterniflora marsh mainly distributed in the high coverage area, andSuaeda salsa marsh mainly distributed in the superior coverage area, which accounted for 45%, 25% and 7.72% of the total vegetation area, respectively.

Key words: grain yield of wheat., fallow period, deep plowing time, soil water storage capacity, dry land