High-accuracy infrared thermography of cotton canopy temperature by unmanned aerial systems (UAS): Evaluating in-season prediction of yield

Authors: SIEGFRIED, JEFFREY - Kansas State University; RAJAN, NITHYA - Texas A&M University; Adams, Curtis; NEELY, HALY - Washington State University; HAGUE, STEVE - Texas A&M University; HARDIN, ROBERT - Texas A&M University; SCHNELL, RONNIE - Texas A&M University; HAN, XIONGZHE - Kangwon National University; THOMASSON, ALEX - Mississippi State University

Submitted to: Smart Agricultural Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/26/2023
Publication Date: 1/7/2024
Citation: Siegfried, J., Rajan, N., Adams, C.B., Neely, H., Hague, S., Hardin, R., Schnell, R., Han, X., Thomasson, A. 2024. High-accuracy infrared thermography of cotton canopy temperature by unmanned aerial systems (UAS): Evaluating in-season prediction of yield. Smart Agricultural Technology. 7(2024). Article 100393. https://doi.org/10.1016/j.atech.2023.100393.
DOI: https://doi.org/10.1016/j.atech.2023.100393

Interpretive Summary: Crop canopy temperature is related to crop and soil water status. New technologies have enabled highly accurate mapping of canopy temperature using thermal cameras onboard unmanned aerial systems (UAS). Using such a system, the objective of this research was to determine relationships between cotton canopy temperature during the growing season and cotton yield and soil moisture. This was accomplished in a two-year field study (2019 and 2020) with three irrigation (0, 40, and 80% ET replacement) and eight cotton variety treatments. Four UAS flights were made per year. Using an approach described within this paper, it was possible to accurately isolate canopy temperature information from the soil background within infrared images using a simple threshold method. There was a close relationship between crop canopy temperature and cotton yields across irrigation levels. When canopy temperature was compared across different cotton varieties, there was evidence that canopy temperature identified subtle varietal differences that were not reflected in yield differences among varieties. Overall, there were linear relationships between canopy temperature and cotton yield that ranged from weak to strong, depending on growing conditions. Canopy temperature was generally correlated with soil volumetric water content, but the relationship was not consistently strong and depended on moisture sensor depth.

Technical Abstract: Canopy temperature in cotton (Gossypium hirsutum) and other crops is related to crop and soil water status. Multiple approaches have been used to measure canopy temperature, depending on the application of the data and available technology. Recent technological advances have made it possible to map canopy temperature using thermal cameras onboard unmanned aerial systems (UAS) at fine spatiotemporal resolution. Using a highly accurate UAS-mounted infrared camera, the objective of this study was to determine relationships of in-season cotton canopy temperature with seed cotton yield and soil moisture. This was accomplished in a two-year field study (2019 and 2020) with three irrigation (0, 40, and 80% ET replacement) and eight cotton variety treatments. Four in-season UAS flights were made per year. Using the approach described herein, it was possible to accurately exclude soil background information in thermal infrared orthomosaics using a simple threshold method without the need for ancillary imagery. This allowed accurate determination of average canopy temperatures across plots. Differences in canopy temperature and seed cotton yields across irrigation levels were largely aligned and, when comparing cotton varieties, there was evidence that canopy temperature identified subtle varietal differences that were not reflected in varietal yield differences. Overall, there were linear relationships between canopy temperature and seed cotton yield that ranged from weak to strong (R2 = 0.39 – 0.72, RMSE = 249.7 – 299.2 kg ha-1), depending on growing conditions. Canopy temperature was generally correlated with soil volumetric water content (VWC), but the relationship was not consistently strong and depended on moisture sensor depth.