Characterizing Wheat Response to Water Limitation Using Multispectral and Thermal Imaging

Effective screening of plant genotypes for their tolerance to abiotic stress is a vital step for crop improvement. Various sensing technologies can be used for developing automated plant phenotyping systems as well as for better control of stress levels imposed on the plants. In this study, seven different wheat genotypes (G1 through G7) were grown under two different water regimes in a greenhouse. Soil moisture was monitored by soil water tension sensors distributed among the experimental plots. A sensor platform with two cameras (a multispectral camera and a thermal infrared camera) was developed to capture top-view images of the wheat plots (once a week) during the course of the experiment. Image processing algorithms were developed to extract wheat growth index (GI) and crop water stress index (CWSI). Ratios of GI and CWSI between the two water treatments were calculated as proxies to assess drought tolerance of the different genotypes. The results showed that GI was correlated with NDVI measured by a GreenSeeker sensor (R2 = 0.65). Among the seven wheat genotypes studied, G5, G6, and G7 exhibited consistently higher GI ratios and lower CWSI ratios, an indication of their superiority in drought tolerance. It was also found that non-uniform environmental conditions in the greenhouse had quite a large influence on wheat growth, which made the characterization and differentiation of phenotypes among various genotypes more challenging. It is concluded that the multispectral and thermal infrared imaging system has potential for phenotypic screening of wheat genotypes for drought tolerance in a semi-controlled environment. Keywords: CWSI, Drought, Image processing, Multispectral imaging, Phenotyping, Thermal infrared imaging.