Submitted to: Cereal Chemistry
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: May 5, 2004
Publication Date: September 1, 2004
Citation: Delwiche, S.R., Hareland, G.A. 2004. Detection of scab damaged hard red spring wheat kernels by near-infrared reflectance. Cereal Chemistry. Vol.81(5):643-649. Interpretive Summary: Fusarium head blight, also known as scab, is a fungal disease that occurs in small grains. Scab in wheat is becoming an increasing problem in the wheat producing regions of the United States and Canada, particularly in seasons of elevated rainfall during grain development, such as in 1993, and most recently in 2003. Scab results in depressed yields and can also adversely affect grain quality. Scabby wheat is a potential health concern, caused by the presence of a mycotoxin known as deoxynivalenol (DON) that is often produced by the fungus. The Federal government has placed advisory levels on DON concentrations for wheat destined for human food or animal feed. For these reasons, both regulator (e.g., USDA-GIPSA) and processor are looking for accurate and objective means for identifying wheat scab in kernels. The current study examined the potential of near-infrared (NIR) reflectance for scab detection in single wheat kernels. Emphasis was placed on identifying two to three wavelengths that could eventually be used in high speed sorting operations. More than 5,000 kernels from commercial releases and breeders lines of hard red spring wheat, equally divided between infected and healthy categories, were examined by single kernel reflectance (1000-1700 nm wavelength range). Using statistical classification techniques, such as linear discriminant analysis and non-parametric (k-nearest-neighbor) classification, we were able to establish a realistic upper level accuracy for NIR-based classification schemes of approximately 97%. An exhaustive search of the most suitable wavelength pairs for the spectral difference, log(1/R at wavelength 1) minus log(1/R at wavelength 2), revealed that the low-wavelength region of a broad carbohydrate absorption band (centered around 1200 nm) was very effective at discriminating between healthy and scab-damaged kernels, with accuracies at about 95%. The achieved accuracy levels demonstrate the potential for the use of NIR in commercial sorting and/or inspection operations for wheat scab.
Technical Abstract: Scab (Fusarium head blight) is a fungal disease that has become increasingly prevalent in North American wheats during the past 15 years. It is of concern to grower, processor, and consumer because of depressed yields, poor flour quality, and the potential for elevated concentrations of the mycotoxin, deoxynivalenol (DON). Both wheat breeder and wheat inspector must currently deal with the assessment of scab in harvested wheat by manual human inspection. The study described herein examined the accuracy of an automated wheat scab inspection system that is based on near-infrared reflectance of individual kernels. Linear discriminant analysis and non-parametric classification (k-nearest-neighbor) models were evaluated for sets that utilized a reduction in 118 wavelengths, as well as a simple difference in 2 wavelengths. Results indicate that a two-wavelength difference discriminant model has the ability to classify normal and scab-damaged kernels with approximately 95% accuracy. The leading slope of the broad 1200-nm carbohydrate absorption band is particularly sensitive for distinguishing these two categories.