Irrigated pinto bean crop stress and yield assessment using ground based low altitude remote sensing technology

Authors: RANJAN, RAKESH - Washington State University; CHANDEL, ABHILASH - University Of Missouri; KHOT, LAV - Washington State University; BAHLOL, HAITHAM - Washington State University; ZHOU, JIANGFENG - Washington State University; Boydston, Rick; Miklas, Phillip - Phil

Submitted to: Information Processing in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/21/2019
Publication Date: 1/29/2019
Citation: Ranjan, R., Chandel, A., Khot, L., Bahlol, H., Zhou, J., Boydston, R.A., Miklas, P.N. 2019. Irrigated pinto bean crop stress and yield assessment using ground based low altitude remote sensing technology. Information Processing in Agriculture. https://doi.org/10.1016/j.inpa.2019.01.005.
DOI: https://doi.org/10.1016/j.inpa.2019.01.005

Interpretive Summary: Pinto beans are the most important dry edible bean market class in the U.S. Our goal is to use plant breeding to develop pinto beans that have higher and more stable yields under stress and nonstress conditions. Pinto bean responses to drought and other stresses can be difficult to measure. High throughput phenotyping of pinto bean performance during the growing season is needed to more accuarately predict yield potential. This study investigates the use of unmanned aerial vehicles equipped with cameras that easily capture multispectral images of thousands of pinto bean breeding lines at a time which may then be useful for predicting health and vigor of individual lines. More than 25 vegetative indices can be generated from a single multispectral image. We identified four of the 25 indices with consistently high correlations between actual performance based on seed yield and estimated performance based on multispectral image data. Further validation of these four new vegetative indices for high throughput phenotyping pinto bean breeding lines for tolerance to stresses is under investigation.

Technical Abstract: This study evaluated the role of multispectral imagery data-based vegetation indices (VIs) for irrigated Pinto bean crop stress and yield assessment. Eight Pinto bean cultivars were grown in fields that had conventional tillage and strip tillage treatments. During production, cultivars received two different levels of irrigation, 52 and 100% of required evapotranspiration. Ground vehicle based 5-band multispectral imaging, at ground sampling distance (GSD) of 4.6 mm/pixel, captured crop vigor variations at three growth stages, i.e., at the early, mid and late growth stage. Multispectral data was analyzed to have 25 VIs that can potentially capture stress traits and the crop yield potential. Principal component analysis, bi-plots analysis and Pearson product moment correlation (r) tests were conducted to identify key VIs and their correlation with abiotic stress at each of the growth stages. Also, the relationship between the VIs that can express vigor variation due to treatments and yield was established in this study. Overall, transformed difference vegetation index (TDVI), nonlinear vegetation index (NLI), modified nonlinear vegetation index (MNLI) and infrared percentage vegetation index (IPVI) were consistent in accounting for stress characterizations and crop yield at all three growth stages (r > 0.63). Moreover, 10 other VIs significantly accounted for crop stress at early and late stages. The TDVI, MNLI, and IPVI had a strong exponential relationship with crop yield in the ranges of 0.51 to 0.56. Overall, the key VIs extracted from such imagery data may be helpful with decision support for precise remedial management practices due to abiotic stressors and in rapid crop yield potential assessment.