Skip to main content
ARS Home » Northeast Area » Kearneysville, West Virginia » Appalachian Fruit Research Laboratory » Innovative Fruit Production, Improvement, and Protection » Research » Publications at this Location » Publication #166027

Title: OPTICAL METHODS FOR DETECTING WATERCORE AND MEALINESS IN APPLES

Author
item BENNEDSEN, BENT
item Peterson, Donald

Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/15/2005
Publication Date: 3/1/2005
Citation: Bennedsen, B.S., Peterson, D.L. 2005. Optical methods for detecting watercore and mealiness in apples. Transactions of the ASAE. PG 803, VOL. 21(5): 803-806 2005 ASAE, ISSN 0883-8542.

Interpretive Summary: Watercore and mealiness are internal disorder in apples, which significantly reduces the value of the fruit, but which can not be detected by external inspection. In order to ensure that defective apples are not passed on to consumers, automatic on-line sorting according to internal quality is required. In this work, spectral reflection in the near infrared area was tested for its ability to detect watercore and mealiness in 'Red Delicious' apples. Watercore was successfully identified in 96% of the cases for apples tested shortly after harvest. Mealiness was detected in 100% of the cases in apples stored about a month after harvest. The system needs to be tested on other varieties. If the performance is equally successful, it could probably be implemented as it is for on-line inspection of apples.

Technical Abstract: Spectral reflection in the wavebands between 600 and 1100 nm was tested for its ability to classify apples according to presence of watercore and mealiness. A model developed by PLS regression using four, 10 nm wide wavebands at 690, 700, 820, and 830 nm separated watercore and normal apples with an error of 4%. It was found that watercore could only be detected shortly after the apples are harvested. Classification of the watercore apples into slight, moderate, and severe watercore may be possible if models are based on a higher number of samples. Mealy apples could be detected from normal and watercore apples with no misclassifications using the reflectance in the wavebands 1020, 1040, 1050, and 1095. Measuring the degree of mealiness based on spectral reflection, was not possible.