Increasing the Rate of Operation of Automatic Quality Classifiers for Agricultural Products - Software and Hardware Decisions

Георгиев, Антон Славчев; Kostadinova, L.; Gabrova, Radoslava Nikolova; Shopov, Nikolay

doi:10.17660/actahortic.2010.858.66

Cited by 3 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The process of the targeted treatment of the resulting primary information for classifying products into quality classes can be divided into the following stages [2,3,5]:…”

Section: Introductionmentioning

confidence: 99%

An Approach for Eliminating Restrictions on the Synthesis of Quality Classifiers of Fruits and Vegetables

Gabrova¹

2014

IJSTS

View full text Add to dashboard Cite

Synthesis of classifiers of agricultural products by quality is related with two main tasks -selection symptoms for object description in new space and choice of method for pattern recognition for objects separation in quality classes. A special feature of agricultural products as objects of classification is the different size of their primary descriptions. This feature imposes restrictions on the choice of symptoms and a method for pattern recognition in the synthesis of the classifier. The work proposes an algorithm for unification of the primary descriptions of classified products using interpolation methods. Different methods of interpolation are compared and the one which provides the simplest algorithm is recommended for use. The algorithm is applied to a virtual extension of experimentally derived primary descriptions of potato tubers. The new extended descriptions are applied to the synthesis of the symptoms and classifier. Simulation testing of the classifier, synthesized with the new and original descriptions of the products was conducted in Matlab. The applicability of the proposed algorithm to unify the descriptions of classified agricultural products is proved. The proposed approach removes the restrictions on the choice of method for the synthesis of symptoms and method for pattern recognition and reduces the number of training set of objects.

show abstract

“…The process of the targeted treatment of the resulting primary information for classifying products into quality classes can be divided into the following stages [2,3,5]:…”

Section: Introductionmentioning

confidence: 99%