Comparison of Remote Sensing Methods for Plant Heights in Agricultural Fields Using Unmanned Aerial Vehicle-Based Structure From Motion

Abstract

Remote sensing using unmanned aerial vehicles (UAVs) and structure from motion (SfM) is useful for the sustainable and cost-effective management of agricultural fields. Ground control points (GCPs) are typically used for the high-precision monitoring of plant height (PH). Additionally, a secondary UAV flight is necessary when off-season images are processed to obtain the ground altitude (GA). In this study, four variables, namely, camera angles, real-time kinematic (RTK), GCPs, and methods for GA, were compared with the predictive performance of maize PH. Linear regression models for PH prediction were validated using training data from different targets on different flights ("different-targets-and-different-flight" cross-validation). PH prediction using UAV-SfM at a camera angle of -60° with RTK, GCPs, and GA obtained from an off-season flight scored a high coefficient of determination and a low mean absolute error (MAE) for validation data (R ² _val = 0.766, MAE = 0.039 m in the vegetative stage; R ² _val = 0.803, MAE = 0.063 m in the reproductive stage). The low-cost case (LC) method, conducted at a camera angle of -60° without RTK, GCPs, or an extra off-season flight, achieved comparable predictive performance (R ² _val = 0.794, MAE = 0.036 m in the vegetative stage; R ² _val = 0.749, MAE = 0.072 m in the reproductive stage), suggesting that this method can achieve low-cost and high-precision PH monitoring.

Keywords: 3D structure analysis; maize; plant height; remote sensing; structure from motion; unmanned aerial vehicle.

FIGURE 1

An overview of the SfM process and three methods for obtaining ground altitude (GA); methods M1, M2, and M3. A DSM is a digital surface model that represents elevation on the 3D model, a DTM is a digital terrain model that represents elevation without plants, and a CHM is a crop height model that represents plant height of crop.

FIGURE 2

An overview of two fields of maize used in this study. Light blue (Field 1) and orange (Field 2) solid lines show the plots. Dash lines show the analysis region of each field. White circles around or in Field 2 show the locations of GCPs. Red rectangles show ROIs on rows (inner ROIs). Blue rectangles show ROIs around the field (outer ROIs).

FIGURE 3

An illustration of the camera angle.

FIGURE 4

An example of noises in point clouds. (A) A point cloud with noises (some points floating over plants). (B) A DSM with noises (an extremely high area near the center). The DSM is shown in grayscale; when the pixel is white, the altitude is high.

FIGURE 5

An example of the process involved in method M3. (A) A digital surface model (DSM) of Field 1. Blue rectangles show ROIs around the field (outer ROIs). (B) Digital terrain models (DTMs) fitted to polynomial surfaces. Blue points show coordinates of the outer ROIs and meshes show the fitted DTMs. “n dim” shows the dimension of the polynomial surface. (C) Crop height models (CHMs) calculated as the difference between DSM and DTMs. These figures were created with the dataset of rep. 1 on Field 1 in the vegetative stage.

FIGURE 6

Scatterplots between measured PH (PH_measured) and PH predicted by the linear regression model from PH_SfM on validation data. “Low-cost case (LC)” is the condition with camera angle: −60°, RTK: unused [−], GCPs: unused [−], method: M3; and “Highest-cost case (HC)” is the condition with camera angle: −60°, RTK: used [+], GCPs: used [+], method: M1. In each set of analysis conditions (LC or HC) and stage (vegetative or reproductive), the nearest to the mean in $R^2{}_{\mathrm{val}}$ was selected from 18 validation cases.

FIGURE 7

A comparison of DSMs with −60° and −90° for the camera angle. The DSMs are shown in grayscale; when the pixel is white, the altitude is high.