Olive classification according to external damage using image analysis

• Published in: Journal of food engineering 2008-08, Vol.87 (3), p.371-379
• Main Authors: Riquelme, M.T., Barreiro, P., Ruiz-Altisent, M., Valero, C.
• Format: Article
• Language: English
• Subjects: anatomy and morphology; Artificial vision; Biological and medical sciences; Calibration; Categories; Classification; color; Defects; External damages; Food engineering; Food industries; Fruit and vegetable industries; fruit quality; Fruits; Fundamental and applied biological sciences. Psychology; General aspects; image analysis; Image processing; Mathematical models; mechanical damage; Olives; product grading; sorting; Sorting fruit; stone fruits; Table olives
• ISSN: 0260-8774; 1873-5770
• DOI: 10.1016/j.jfoodeng.2007.12.018

The external appearance of an olive’s skin is the most decisive factor in determining its quality as a fruit. This work tries to establish a hierarchical model based on the features extracted from images of olives reflecting their external defects. Seven commercial categories of olives, established by product experts, were used: undamaged olives, mussel-scale or ‘serpeta’, hail-damaged or ‘granizo’, mill or ‘rehús’, wrinkled olive or ‘agostado’, purple olive and undefined-damage or ‘molestado’. The original images were processed using segmentation, colour parameters and morphological features of the defects and the whole fruits. The application of three consecutive discriminant analyses resulted in the correct classification of 97% and 75% of olives during calibration and validation, respectively. However the correct classification percentages vary greatly depending on the categories, ranging 80–100% during calibration and 38–100% during validation.