| Popis: |
Upland cotton (Gossypium hirsutum L.) requires extensive crop management using chemicals such as defoliants. Since they are generally spread in uniform applications across a field, defoliants tend to be overapplied in some areas. This leads to increased chemical runoff and decreased profit for the farmer. To help improve this process, remotely sensed data acquired via unmanned aircraft systems (UAS) were used to develop variable rate prescription maps to drive cotton defoliant application. Flights were performed on a weekly basis over a 10-ha cotton field in Goldsboro, NC. Data collected via a Parrot Sequoia multispectral sensor and a Zenmuse X3 multispectral sensor were used to create normalized difference vegetation index (NDVI) maps to monitor plant health and digital surface models (DSMs) to measure plant height. Using a combination of the relative high and low values for the two crop indicators, variable rate prescription maps were developed for two chemicals, Folex, the main defoliant, and Dropp, a regrowth inhibitor. The field was divided into eight-row swaths which received either variable or fixed rate defoliant applications. Cotton lint yield data was collected at harvest to detect any negative effects from the variable rate application. Using the variable rate prescription maps reduced the amount of Folex used by 10.3% and the amount of Dropp used by 17.6%. There was not a significant difference in cotton lint yield between the different defoliant treatments. Based on these results, UAS have the potential to make cotton defoliation more efficient and increase profit for the farmer. |
