Classification modeling based on surface porosity for the grading of natural cork stoppers for quality wines

Abstract

The natural cork stoppers are commercially graded into quality classes according with the homogeneity of the external surface. The underlying criteria for this classification are subjective without quantified criteria and standards defined by cork industry or consumers. Image analysis was applied to premium, good and standard quality classes to characterize the surface of the cork stoppers and stepwise discriminant analysis (SDA) was used to build predictive classification models. The final goal is to analyze the contribution of each porosity feature and propose an algorithm for cork stoppers quality class classification. This study provides the knowledge based on a large sampling to an accurate grading of natural cork stoppers. In average all the models presented accuracy in relation to the commercial classification over 68% with a higher mismatch in the mid-quality range. Color showed an important discriminating power, increasing the accuracy in 10%. The main discriminant features were porosity coefficient and color variables, calculated for the lateral surface. A quality classification algorithm was presented based on a simplified model with an accuracy of 75%. The classification based on color vision systems can ensure improved quality class uniformity and a higher transparency in trade.

Cork and Cork Stoppers: Quality and Performance

“…Given its importance, the cork plank thickness is normalized by caliper classes (NP 298:1993 and ISO 1219:1998). Having the suitable thickness, the plank quality is given mostly by the cork porosity [20,21] that also determines the suitability and quality yields of the products [22][23][24]. Porosity quantification by surface image analysis is the basis of cork stoppers classification into quality grades [24,25].…”

Cork, a Natural Choice to Wine?

“…However, despite being only one piece, they can be graded in different categories. The external surface homogeneity/heterogeneity of the natural cork stoppers will determine the commercial grade and the quality classes [ 31 ]. The homogeneity of the cork surface is given by the absence of voids or defects (the presence of lenticular channels), which is recognised as the porosity of the cork [ 32 ].…”

Feature selection for surface defect classification of extruded aluminum profiles

“…A further investigation by Petrov et al, using part of the same dataset and additional features such as Entropy, tested the performance of self-organizing maps with success rates up to 88 % on PCA and 98 % on LDA [18]. Concerning cork classification, Oliveira et al used stepwise discriminant analysis to build predictive classification models to characterize the surface of cork stoppers and cluster them into three quality classes, analyzing the contribution of each porosity feature in the classification [13]. A difficult task was for Tsai and Chao to detect defects in sputtered glass surfaces, since the nature of those surfaces is random and anomalies are observed [25].…”