Detection of Sour Skin Diseases in Vidalia Sweet Onions Using Near-infrared Hyperspectral Imaging

Abstract

Vidalia sweet onion is an important specialty crop grown in southeast area of Georgia, which accounts for approximate 13% of the vegetable farm gate value in the state. However, Vidalia sweet onions are more prone to rotting and sprouting due to lacking some pungent compounds found in other onion varieties. Sour skin, a common onion disease caused by the bacterium Burkholderia cepacia, can result in significant losses during storage. Moreover, some strains of the bacterium also may result in pulmonary infections in humans and have increasingly been associated as a leading cause of nosocomial infections. Therefore, it is critical to prevent onions infected by sour skin from entering in storage rooms or being shipped to fresh vegetable markets. The objective of this research was to investigate using a liquid crystal tunable filter (LCTF) based hyperspectral imaging (HSI) system to identify sour skin in sweet onions. A number of preliminary tests were conducted to study transmittance and reflectance spectral characteristics of sour skin-infected sweet onions using this HSI system. Transmittance tests suggested that the light source used in this study could penetrate three layers with good transmittance. The reflectance tests proved that a sour skin infected region was darker than healthy flesh region, and the best contrast was in the spectral region of 1200-1300 nm. In addition, a better contrast between Vidalia sweet onion surface dry layer and inner fresh layers in HSI images was found in the spectral region of 1400-1500 nm. Mean reflectance spectra from a sour skin infected area indicated that there was a significant difference sensed by HSI reflectance spectra in the spectral region of 1150-1280 nm when the onion was stored 3 days after being inoculated. This work provides a foundation for future study of detecting sour skin disease in Vidalia sweet onions via hyperspectral or multispectral imaging tools.