Automated Nondestructive Detection of Internal Insect Infestation of Wheat Kernels by Using Near-Infrared Reflectance Spectroscopy

Abstract

Wheat kernels infested internally with larvae of 3 primary insect pests of grain, the rice weevil, Sitophilus oryzae (L.); the lesser grain borer, Rhyzopertha dominica (F.); and the Angoumois grain moth, Sitotroga cerealella (Olivier), were scanned with a near-infrared spectrometer incorporated into a single kernel characterization system to determine differences in absorption due to the presence of larvae. The single kernel characterization system delivers kernels into the spectrometer viewing area at the rate of 1 per 4 s. We were able to differentiate uninfested kernels from kernels infested with larvae of all 3 species by using this automated system. Moisture content, protein content, or wheat class did not affect classification accuracy. The calibration included spectral characteristics in the wavelength ranges of 1,000-1,350 and 1,500-1,680 nm. Larval size was a factor in the sensitivity of the system, with 3rd and 4th instars rice weevil being detected with 95% confidence. In contrast to many other procedures used to detect internal insect infestations in grain, this system could be incorporated into the current grain inspection process and provide the grain industry with quantitative data on internal insect infestations in wheat.