A New Method for Winter Wheat Mapping Based on Spectral Reconstruction Technology

Timely and accurate estimation of the winter wheat planting area its spatial distribution is essential for implementation crop growth monitoring yield estimation, hence development national agricultural production food security. In remotely sensed mapping based on spectral similarity, reference curve obtained by averaging multiple standard curves, which limits accuracy. We propose a reconstruction method singular value decomposition (SR-SVD) unique characteristics crops. Using Sentinel-2 A/B satellite data, we tested SR-SVD in Puyang County, Shenzhou City, China. Performance was increased, with optimal overall accuracy Kappa County City were 99.52% 0.99, 98.26% 0.97, respectively. selected angle mapper (SAM) Euclidean Distance (ED) as similarity measures. Compared to methods, significantly improves accuracy, it avoids excessive extraction, can identify more detailed information, advantageous distinguishing non-winter pixels. Three commonly used supervised classification support vector machine (SVM), maximum likelihood (ML), minimum distance (MD) comparison. Results indicate that has highest greatly reduces number misidentified Therefore, achieve high-precision provide technical regional structure information.