Sampling Survey Method of Wheat Ear Number Based on UAV Images and Density Map Regression Algorithm

The number of wheat ears is one the most important factors in yield composition. Rapid and accurate assessment ear great importance for predicting grain food security-related early warning signal generation. current counting methods rely on manual surveys, which are time-consuming, laborious, inefficient inaccurate. Existing non-destructive detection techniques mostly applied to near-ground images difficult apply large-scale monitoring. In this study, we proposed a sampling survey method based unmanned aerial vehicle (UAV). Firstly, small UAV were acquired five-point mode. Secondly, an adaptive Gaussian kernel size was used generate ground truth density map. Thirdly, map regression network (DM-Net) constructed optimized. Finally, designed overlapping area sub-images solve repeated caused by image segmentation. MAE MSE model 9.01 11.85, respectively. We compared paper with method. results showed that RMSE MAPE NM13 18.95 × 104/hm2 3.37%, respectively, YFM4, 13.65 2.94%, This study enables investigation large area, can provide favorable support estimation.