Estimating yield-contributing physiological parameters of cotton using UAV-based imagery

• DOI: 10.3389/fpls.2023.1248152
• Publication Date: 2023-09-19T20:32:36Z
• Authors: Amrit Pokhrel; Simerjeet Virk; John L. Snider; George Vellidis; Lavesta C. Hand; Henry Y. Sintim; Ved Parkash; Devendra P. Chalise; Joshua M. Lee; Coleman Byers
• Keywords: Lint; Tifton; Growing season; Photosynthetically active radiation; Petiole (insect anatomy)

Lint yield in cotton is governed by light intercepted the canopy (IPAR), radiation use efficiency (RUE), and harvest index (HI). However, conventional methods of measuring these yield-governing physiological parameters are labor-intensive, time-consuming requires destructive sampling. This study aimed to explore low-cost high-resolution UAV-based RGB multispectral imagery 1) estimate fraction IPAR (IPAR f ), RUE, biomass throughout season, 2) lint using fiber (CFI), 3) determine potential models for estimating HI. An experiment was conducted during 2021 2022 growing seasons Tifton, Georgia, USA randomized complete block design with five different nitrogen treatments. Different treatments were applied generate substantial variability development yield. UAV collected bi-weekly along interception measurements 20 vegetation indices (VIs) computed from imagery. Generalized linear regression performed develop VIs degree days (GDDs). The had R 2 values ranging 0.66 0.90, based on RVI RECI explained highest variation (93%) cross-validation. Similarly, above-ground best estimated MSAVI OSAVI. Estimation RUE actual measurement RVI-based model able explain 84% RUE. CFI strong relationship (R = 0.69) machine harvested HI CFI-based MSAVI-based 40 49% measured season. developed yield-contributing showed low performance, having greater prediction accuracy. Future studies accurate estimation suggested precise prediction. first attempt its kind results can be used expand improve research predicting functional drivers cotton.