UAS-derived pest management solution to sorghum crop production
Conference Paper
Overview
Overview
abstract
The flexibility of image acquisition afforded by unmanned aircraft systems (UAS) equipped with tnulpsectr al sensors provide an effective tool for managing aphid infestation of sorghum. Machine learning algorithms and statistical analysis of spectral and structural information from remotely sensed imagery provide enhanced insights into crop damage from aphids. High temporal repeatability during the crop growth cycle afford the opportunity for improved management of infestations. In this work, thelstudy area is a two-acre sorghum fie d experiment on economic spray thresholds for sugarcane aphid infestation at the Texas A&M AgriLife research facility in Corpus Christi, TX USA. Imagery from a fixed-wing UAS, eBee, equipped with a three-band NlR-R-G sensor was obtained on a biweekly to weekly basis over the summer 2016 growing season, imagery was also obtained using quad-copter. Phantom 3, flying at 25m elevation, in order to obtain high resolution images to create a point cloud modeling the crop. The field was comprised of an early crop and a late crop sharing an equal portion of field. Data was also collected manually to verify remotely sensed estn tes. The imagery was processed to derive orthoretif d reflectance maps from which Normalized Difference Vegetation index (NDVl) and 3D point cloud data crop structure were derived.
