Hyperspectral Fluorescence Imaging Detection System for Black Walnut Shell Fragments Recognition and Removal

Non-Tech Summary: The black walnut grows throughout the central and eastern United States. It's rich, tasty flavor is a high-valued food additive. Although 15.4 acres of black walnuts yield approximately 4 billon pounds of raw nuts annually, only about 20 million pounds are commercially processed. With low economic incentives and small processing capacity, growers lack motivation to harvest. However, the growers can potentially enjoy great economic gain if these black walnuts can efficiently sorted. The USDA's very strict standards allow no more than 0.05% shell or foreign material by weight in any grade of shelled nuts. With very similar physical properties between the shells and meat, an effective method of automatically differentiating the shell from the meat needs developing, particularly for black walnut processing. This research will develop a hyperspectral fluorescence imaging technology to solve the challenging problem of distinguishing between nut fragments from nut meat which are much similar in physical characteristics. This technique utilizes the fluorescence emitted light responses under hyperspectral wavebands and high-dimensional imaging algorithms to identify the shell fragments and nut meats. Further, this automated dynamic system technology will enable the removal of nut shells from the nut meat non-destructively for quality and safety nut products. This research will lead to a valuable technology toward an initial prototype which allows growers to recoup economic benefit from their black walnuts.