Sagan, Vasit (ORCID:0000000343752096)

DOE OSTI2023

Early Detection of Drought Stress in Durum Wheat Using Hyperspectral Imaging and Photosystem Sensing

Wheat, being the third largest U.S. crop and the principal food grain, faces significant risks from climate extremes such as drought. This necessitates identifying and developing methods for early water-stress detection to prevent yield loss and improve water-use efficiency. This study investigates the potential of hyperspectral imaging to detect the early stages of drought stress in wheat. The goal is to utilize this technology as a tool for screening and selecting drought-tolerant wheat genotypes in breeding programs. Additionally, this research aims to systematically evaluate the effectiveness of various existing sensors and methods for detecting early stages of water stress. The experiment was conducted in a durum wheat experimental field trial in Maricopa, Arizona, in the spring of 2019 and included well-watered and water-limited treatments of a panel of 224 replicated durum wheat genotypes. Spectral indices derived from hyperspectral imagery were compared against other plant-level indicators of water stress such as Photosystem II (PSII) and relative water content (RWC) data derived from proximal sensors. Our findings showed a 12% drop in photosynthetic activity in the most affected genotypes when compared to the least affected. The Leaf Water Vegetation Index 1 (LWVI1) highlighted differences between drought-resistant and drought-susceptible genotypes. Drought-resistant genotypes retained 43.36% more water in leaves under well-watered conditions compared to water-limited conditions, while drought-susceptible genotypes retained only 15.69% more. The LWVI1 and LWVI2 indices, aligned with the RWC measurements, revealed a strong inverse correlation in the susceptible genotypes, underscoring their heightened sensitivity to water stress in earlier stages. Several genotypes previously classified based on their drought resistance showed spectral indices deviating from expectations. Results from this research can aid farmers in improving crop yields by informing early management practices. Moreover, this research offers wheat breeders insights into the selection of drought-tolerant genotypes, a requirement that is becoming increasingly important as weather patterns continue to change.

Roy, Bishal (ORCID:0000000199125505)↗

DOE OSTI2022

Data-Driven Artificial Intelligence for Calibration of Hyperspectral Big Data

Not Available

42 ENGINEERING↗

DOE OSTI2020

Quantifying Leaf Chlorophyll Concentration of Sorghum from Hyperspectral Data Using Derivative Calculus and Machine Learning

Leaf chlorophyll concentration (LCC) is an important indicator of plant health, vigor, physiological status, productivity, and nutrient deficiencies. Hyperspectral spectroscopy at leaf level has been widely used to estimate LCC accurately and non-destructively. This study utilized leaf-level hyperspectral data with derivative calculus and machine learning to estimate LCC of sorghum. We calculated fractional derivative (FD) orders starting from 0.2 to 2.0 with 0.2 order increments. Additionally, 43 common vegetation indices (VIs) were calculated from leaf spectral reflectance factor to make comparisons with reflectance-based data. Within the modeling pipeline, three feature selection methods were assessed: Pearson’s correlation coefficient (PCC), partial least squares based variable importance in the projection (VIP), and random forest-based mean decrease impurity (MDI). Finally, we used partial least squares regression (PLSR), random forest regression (RFR), support vector regression (SVR), and extreme learning regression (ELR) to estimate the LCC of sorghum. Results showed that: (1) increasing derivative order can show improved model performance until certain order for reflectance-based analysis; however, it is inconclusive to state that a particular order is optimal for estimating LCC of sorghum; (2) VI-based modeling outperformed derivative augmented reflectance factor-based modeling; (3) mean decrease impurity was found effective in selecting sensitive features from large feature space (reflectance-based analysis), whereas simple Pearson’s correlation coefficient worked better with smaller feature space (VI-based analysis); and (4) SVR outperformed all other models within reflectance-based analysis; alternatively, ELR with VIs from original reflectance yielded slightly better results compared to all other models.

47 OTHER INSTRUMENTATION↗