The USDA/ARS fruit postharvest technology program in the Agricultural Engineering Department at Michigan State University is developing new sensing technologies to help the fruit industry deliver better, more consistent fruit to the consumer. The research program is fully funded by USDA, with additional funds from the fruit industry, and is in partnership with the AE Department at MSU to address the priority needs of the fruit industry in Michigan and the nation. The mission of this research program is to develop new knowledge, methods and systems for assessing, retaining, and enhancing postharvest quality of tree fruits, particularly apples and cherries. Current research is focused on developing nondestructive optical sensing technologies for assessing, grading and sorting fruit for external and internal quality characteristics to provide accurate, rapid, and reproducible information on fruit.

Packing House click here for larger view

 

Currently, most packinghouses sort fruit for size and color, but technologies for sorting internal quality are not available. A great looking fruit cannot guarantee great taste. Fruit growers, packers, and USDA inspectors still rely on subjective means, such as touching and tasting, and other destructive testing methods to gauge fruit internal quality. Once the fruit is tested, it has to be discarded. These grading and inspection practices cannot guarantee the quality of individual fruit delivered to the consumer because there is great variation in texture (firmness, crunchiness, juiciness, etc.) and flavor (sweetness, sourness, etc.) from the same lot of apple fruit. Poor, inconsistent fruit quality has turned some consumers away to look for other food products, causing the industry to lose market share and competitiveness.

We are now investigating several novel techniques to look beyond fruit skin to detect internal quality of fruit. Hyperspectral imaging is such one technique that is being investigated. The technique combines conventional two-dimensional digital imagery with spectroscopy to analyze various wavelengths in the visible and near-infrared (non-visible) region to ascertain certain minor and/or subtle features in an object - whether it is terrain, a camouflaged armored vehicle, or an apple. A prototype hyperspectral imaging system has been developed that measures light bouncing back from the fruit after it interacts with fruit tissue. The hyperspectral images contain useful information about light absorption and scattering in the fruit. Computer algorithms incorporating artificial neural networks features are being developed to analyze the hyperspectral images of apples for predicting firmness and sugar content. Our initial study showed good results in predicting apple firmness and sugar content and the technique has the potential for grading and sorting fruit for these quality attributes. Our next goal is to optimize the system's parameters and computer algorithms to further improve predictions and then integrate hardware and software for rapid detection of fruit firmness and sugar content. We hope that the technique will be eventually transferred to the industry for sorting and grading apples as well as other fruits. So some day in the future when you shop fresh produce in grocery stores, you can get the apples you really want.