Infrared chemical imaging of germinated wheat: early nondestructive detection and microspectroscopic imaging of kernel thin cross sections in Situ

Abstract

During germination, biochemical changes occur in the wheat kernel by stimulation of enzymes and hormones, and the seed reserves are mobilized. Infrared microspectroscopy and imaging enables a localized chemical inventory, upon germination, to study the process. Frozen sections of germinated wheat mounted onto BaF[subscript]2 were mapped to produce functional group images for comparison with corresponding sections of ungerminated kernels. Relative functional group populations in the scutellum and embryonic axis were assessed before and after germination. An average 23% reduction in lipid to protein ratio was observed in the scutellum based on the comparison of 53,733 spectra. As a result of the early germination process, lipid in the scutellum was depleted to provide energy for the growing embryo. Germination of the kernels while in the field before harvest due to high humidity is known as preharvest sprouting. Preharvest sprouting has detrimental effects on the end use quality of the wheat (sprout damage) and cause economic loses. Tolerance to preharvest sprouting is highly desirable. To assist breeding program, a nondestructive near-IR chemical imaging method has been developed to test new lines for resistance to preharvest sprouting. The higher sensitivity of subsurface chemical imaging, compared with visual detection, alpha amylase determination, or viscosity testing, permits germination detection at early stages. A near-IR chemical imaging system with an InGaAs focal plane array (FPA) detector in the 1100 nm-1700 nm range was used. Kernels from six different cultivars, including HRW and HWW wheat, were exposed to moist conditions for 6, 12, 24, 36, and 48 hours. Images of each 90 kernel group were examined; kernels exposed to moisture for 36 hours were compared with kernels treated for 3 hours as a control. Each kernel was classified as sprouted or not sprouted with the criteria of log 1/R intensity at select wavelengths or select factors of principle component analysis (PCA) treatment of reflectance intensity data. Imaging wavelength range was expanded beyond 1700 nm to 2400 nm with the use of InSb FPA. Study for the potential for unsupervised determination in nondestructive near-IR imaging with detection wavelengths 1200-2400 is ongoing. Some preliminary results presented are encouraging.