Multivariate Optical Element Platform for Post-Harvest Inspection

Over the past two decades, Multispectral Imaging (MSI) and Hyperspectral Imaging (HSI) technologies have been fielded as rapid, non-destructive quality and safety inspection systems for a vast range of food products. Specifically, MSI and HSI systems enable the evaluation of overall food quality via classification and grading, defect and disease detection as well as visualization of chemical distributions. Today's MSI and HSI systems are expensive, complicated and not broadly deployed, except by the largest growers and processors. MSI systems utilize a small set of discrete wavelength measurements while HSI systems may utilize more than 100 discrete wavelength measurements for evaluating food product quality. Ultimately, these systems may be employed instead of human inspectors or wet chemical methods for the automatic grading and nutrition determination of food products. HSI system configurations exist in visible and short-wave infrared (SWIR) spectral regions which employ powerful multivariate analysis detection algorithms. While these imaging systems can provide sensitive and specific detections of food quality targets, they are typically costly to field, operate, and support. As a result, these systems are not broadly deployed, except by the largest growers and processors. In order to achieve real-time operation, some degree of freedom, such as the number of spectral bands, image definition, or a number of food quality targets being detected is compromised. HSI systems also generate a "big data" problem in which the backend data handling and processing requirements become an ever growing problem.One specific, post-inspection agricultural challenge is correctly sorting peaches based upon size, dissolved solid content (sugar) and ripeness. Peaches are first graded by ripeness in the field with a handheld sensor that measures the chlorophyll content in the skin. Next, a high-speed sorter measures and sorts the fruit based upon physical size. SWIR HSI technology is employed to identify subsurface defects in parallel to the process. A Brix measurement is made from batch to batch to ensure that a minimum level of sugar content is being achieved. Ultimately, a uniformly sized set of peaches is sorted with varying levels of ripeness and sugar content leading to unanticipated product quality at the final point of purchase. An ideal imaging system would grade incoming peaches according to size, dissolved solid content (sugar) and ripeness which impacts multiple levels of the consumer chain:Farmers reduce labor costs and improve delivery of high quality, ripened peaches to retailersRetailers receive more uniform products for merchandising, which can help reduce waste and costsConsumers receive "better tasting" products, which leads to more demand for peachesMultivariate Optical Elements (MOEs) are thin-film devices that encode a broadband, spectroscopic pattern allowing a simple broadband detector to generate a highly sensitive and specific detection of a target analyte. MOE filter sets are capable of sensing an orthogonal projection of the original sparse spectroscopic space enabling a small set of MOEs to discriminate a multitude of target analytes. CIRTEMO is working with commercial partners to develop a MOE sensor for the detection of peach analyte targets such as ripeness and Brix content. This MOE-based, real-time HSI sensor will exhibit superior sensitivity and specificity as compared to currently fielded MSI and HSI systems in addition to reducing the overall low Size, Weight, Power and Cost (SWAPc) package. By reducing the costs and complexity of today's HSI systems, CIRTEMO will enable broader deployments of HSI technology which will benefit both large and small farmers and processors by reducing waste and labor/operating costs.