| Plant height and leaf area index (LAI) are important agronomic traits of crops that reflect their growth and development to some extent. In this study, we aimed to explore the reliability of winter wheat plant height extraction based on UAV visible remote sensing and assess the accuracy of LAI estimation using plant height and visible vegetation index. We obtained the plant height and visible vegetation index, and used four modeling methods (stepwise regression, partial least squares, random forest, and artificial neural network) to establish LAI estimation models. Subsequently, we evaluated the accuracy of plant height extraction and LAI estimation. The results showed that: (1) Extracted plant height Hc was well fitted with measured plant height Hd (R=0.894, RMSE = 6.695, NRMSE = 9.63%), indicating good performance in plant height extraction; (2) In contrast to use of the visible light vegetation index only, the LAI estimation model constructed based on plant height and visible light vegetation index had higher accuracy. In addition, the random forest was the optimal modeling method, when the number of decision trees was 50, the model estimated the best effect (R = 0.809, RMSE = 0.497, NRMSE = 13.85%, RPD = 2.336). Using UAV visible light remote sensing method, it has a high possibility to extract the plant height and estimate the LAI of winter wheat efficiently, accurately and without damage. The results can provide a reference for remote sensing based monitoring of agricultural conditions. |
