Estimation of Nitrogen in Rice Crops from UAV-Captured Images

Abstract

Leaf nitrogen (N) directly correlates to chlorophyll production, affecting crop growth and yield. Farmers use soil plant analysis development (SPAD) devices to calculate the amount of chlorophyll present in plants. However, monitoring large-scale crops using SPAD is prohibitively time-consuming and demanding. This paper presents an unmanned aerial vehicle (UAV) solution for estimating leaf N content in rice crops, from multispectral imagery. Our contribution is twofold: (i) a novel trajectory control strategy to reduce the angular wind-induced perturbations that affect image sampling accuracy during UAV flight, and (ii) machine learning models to estimate the canopy N via vegetation indices (VIs) obtained from the aerial imagery. This approach integrates an image processing algorithm using the GrabCut segmentation method with a guided filtering refinement process, to calculate the VIs according to the plots of interest. Three machine learning methods based on multivariable linear regressions (MLR), support vector machines (SVM), and neural networks(NN), were applied and compared through the entire phonological cycle of the crop: vegetative (V), reproductive (R), and ripening (Ri). Correlations were obtained by comparing our methods against an assembled ground-truth of SPAD measurements. The higher N correlations were achieved with NN: 0.98 (V), 0.94 (R), and 0.89 (Ri). We claim that the proposed UAV stabilization control algorithm significantly improves on the N-to-SPAD correlations by minimizing wind perturbations in real-time and reducing the need for offline image corrections.

UAVs for Vegetation Monitoring: Overview and Recent Scientific Contributions

“…Therefore, it is important to ensure a robust stabilization system that provides high-quality image data and, thus, improves the precision of the image registration algorithms. Colorado et al (2020) [30] developed a novel UAV stabilization control called BS+DAF (backstepping + desired angular acceleration function) to reduce angular wind-induced disturbances. The BS+DAF consisted of a nonlinear trajectory-tracking controller that adds aerodynamics information to control yaw, producing roll and pitch acceleration commands to reduce sudden angular acceleration changes due to external perturbations as those caused by wind.…”

“…A ML approach was also employed by Colorado et al (2020) [30] to monitor canopy N in rice crops with multispectral UAV imagery. The approach was evaluated during the dry season at three rice growth stages: vegetative, reproductive, and ripening.…”

Novel Feature-Extraction Methods for the Estimation of Above-Ground Biomass in Rice Crops

“…For instance, authors in [15] conducted a comprehensive survey to identify which VIs were suitable for estimating rice biomass as a function of the growth stages of the crop. Furthermore, in [15,16], crop features were extracted by using classical k-means binary clustering for segmentation, where linear multi-variable regression models were applied to each crop stage independently for the biomass characterization. In [3], a more sophisticated method named GFKuts was presented, aimed at combining a Monte Carlo K-means with the well-known GrabCut segmentation method [17,18].…”

“…Learn the graph by using the prior of smoothness [16] Apply blue-noise sampling [15] over the smoothed graph…”

“…NARMAX (Nonlinear Autoregressive Moving Average with Exogenous Inputs) methods provide models that are transparent, and easily solve many problems such as contour errors in CNC machines, forecasting, network traffic, and prediction of daily solar radiation [38][39][40][41]. The NARMAX model is defined in Equation (16). Where y(k) is the output system, u(k) is the input system and e(k) is the noise sequence.…”