Submitted to: Proceedings of SPIE
Publication Type: Proceedings
Publication Acceptance Date: May 20, 2010
Publication Date: June 1, 2010
Citation: Delwiche, S.R., Kim, M.S., Dong, Y. 2010. Damage and quality assessment in wheat by NIR hyperspectral imaging. Proceedings of SPIE. 767607:8. Interpretive Summary: In the United States and elsewhere, wheat plants during their flowering and seed head development stages are often infected with a fungal disease known as Fusarium head blight. This disease causes depressed yields, lower quality flour, and worst of all, the potential for accompaniment by deoxynivalenol, a metabolite of the fungus that is moderately toxic to humans and animals. Government and private inspection procedures include a component that requires visual inspection for Fusarium damage. We explored the possibility of using an alternative procedure, near-infrared spectroscopic (hyperspectral) imaging, to assess Fusarium damage. We found that images from as few as two discrete wavelengths could collectively distinguish damaged and normal kernels. These findings, when coupled with similar image analysis procedures for other forms of damage (heat-, frost-, other molds, and insect-), will be of potential benefit to official inspection offices and commercial processors.
Technical Abstract: Fusarium head blight is a fungal disease that affects the world's small grains, such as wheat and barley. Attacking the spikelets during development, the fungus causes a reduction of yield and grain of poorer processing quality. It also is a health concern because of the secondary metabolite, deoxynivalenol, which often accompanies the fungus. While chemical methods exist to measure the concentration of the mycotoxin and manual visual inspection is used to ascertain the level of Fusarium damage, research has been active in developing fast, optically based techniques that can assess this form of damage. In the current study a near-infrared (1000-1700 nm) hyperspectral image system was assembled and applied to Fusarium-damaged kernel recognition. With anticipation of an eventual multispectral imaging system design, 5 wavelengths were manually selected from a pool of 146 images as the most promising, such that when combined in pairs or triplets, Fusarium damage could be identified. We present the results of two pairs of wavelengths [(1199, 1474 nm) and (1315, 1474 nm)] whose reflectance values produced adequate separation of kernels of healthy appearance (i.e., asymptomatic condition) from kernels possessing Fusarium damage.