Crop height monitoring with digital imagery from Unmanned Aerial System (UAS) Academic Article uri icon

abstract

  • © 2017 Crop height is a very important attribute to assess overall crop condition, irrigation, and estimation of terminal yield. In this study, a novel method to monitor crop height of Sorghum (Sorghum bicolor) using an Unmanned Aerial System (UAS) is proposed. UAS data were acquired seven times over the growing season and each aerial acquisition included over 200 images with significant image overlap at an altitude of 50 m above ground. Ortho-mosaic image and 3D point cloud were generated by applying the Structure from Motion (SfM) algorithm to the images. Ground control points (GCPs) were installed around the study area and they were surveyed using a real time kinematic (RTK) GPS unit for accurate geo-referencing of the geospatial data products. A Digital Terrain Model (DTM) and Digital Surface Model (DSM) were generated from the 3D point cloud data, and a Crop Height Model (CHM) was then created by subtracting DTM from DSM. Uniform crop grids along the center line of each variety were defined for further processing. The maximum CHM value within each individual grid was taken to represent crop height of the grid, and average of all grid heights over the whole area of each variety was calculated as crop height of individual variety. These measurements were compared with manual crop height measurements. Root Mean Square Error (RMSE) between field measurements and the proposed approach was 0.33 m. In addition, the height estimates from both field measurement and the proposed approach could be used to derive a growth curve by fitting a sigmoidal curve. The residual RMSEs between the observed and predicted value of the curve established from UAS and field measurements were calculated as 0.05 m and 0.1 m, respectively. The growth curve results showed that the proposed approach indicated less RMSE and generated more reliable growth curves for monitoring sorghum height.

author list (cited authors)

  • Chang, A., Jung, J., Maeda, M. M., & Landivar, J.

publication date

  • January 1, 2017 11:11 AM