Non-invasive estimation of relative water content in soybean leaves using infrared thermography

Infrared thermography is a useful technology for examining water status in terrestrial vegetation. This research was focused on assessing the water status of soybean plants [Glycine max (L.) Merrill] using high resolution thermal infrared images. The plants were subjected to a range of moisture stress treatments in order to evaluate the water content in sampled leaves. The plants were irrigated with 8 different treatment levels [control (i.e., fully irrigated) and 1 to 7 days of water being withheld]. One specific trifoliate was segmented from each of the thermal images for every plant sample, and both mean temperature and Crop Water Stress Index (CWSI) were computed for each plant. Leaf discs were taken from the same trifoliate to gravimetrically measure relative water content (RWC). RWC had statistically significant correlation coefficients with both CWSI (r = -0.92, n = 56; p < 0.001) and raw mean temperature (r = -0.84, n = 56; p < 0.001). Two separate regression models were developed to predict RWC using mean raw trifoliate temperature and CWSI. Our results document that a CWSI-based regression model was better in predicting RWC than a model based on mean raw trifoliate temperature.