基于多时相Sentinel-2卫星影像的冬小麦面积提取OACSTPCD

Timely and accurate extraction of winter wheat planting information is of great research significance for remote sensing monitoring of winter wheat growth.In this study,the Sentinel-2 satellite remote sensing images of winter wheat during overwintering stage(2021-12-04),flowering stage(2022-04-08)and milk ripening stage(2022-05-03)in Yuhang District were used as data sources.The winter wheat planting area was extracted by the maximum likelihood classification,support vector machine,normalized difference vegetation index(NDVI)addition and subtraction synthetic operation methods,respectively.Combining the field survey data with the measured planting area of winter wheat,the accuracy of the results extracted by different classification methods were evaluated.The results showed that using threshold value of NDVI during overwintering stage to mask evergreen vegetation areas(tea garden,woodland)and performing addition operations on the NDVI values of non-evergreen vegetation areas(buildings,water bodies,winter wheat)during flowering and milk ripening stages was the optimum method for extracting the planting area of winter wheat in Yuhang District,with an area accuracy of 91.96％.The results indicated that multi-temporal remote sensing images combined with the phenological characteristics of vegetation and typical land types could obtain high-precision planting area extraction of winter wheat.