It is of great significance to accurately grasp the soybean planting areas for food security in China. In this study,the full temporal images of soybean growth cycle were selected to carry out high-accuracy extraction of soybean planting information. Through the development of this study,the following conclusions were obtained:through multi-temporal spectral comparison,the degree of soybean spectral discrimination on August 7th was the highest;five indexes including NDVI,DVI,NDWI,NDre1 and NDre2,were selected through calculation and comparison of multi-source indexes. Through object-oriented classification,all objects are completely extracted;the extracted plot boundary is in good agreement with the actual field ridge trend,and a variety of index constraints are add-ed. Crops are better differentiated from other vegetation,while forest land,grassland,slope protection shrubs and grasses have less in-fluences on crop extraction. And through post-processing,more complete extraction of all soybean planting information is used to veri-fy the extraction accuracy of soybean planting area extracted by remote sensing images. The final correlation coefficient of remote sens-ing interpretation is 0.7695 and the extraction accuracy is relatively high.
维普
万方数据
