Image Processing and Neuro-Fuzzy Computing for Cork Quality Classification

In this paper we solve a classification problem existing in the cork industry: the cork stopper/disk quality classification. Cork is a material that can be mostly obtained in the occidental shores of the Mediterranean Sea and its production has a great importance in those areas. Due to cork is a natural and heterogeneous material its automatic quality detection (usually, seven different quality classes exist) is difficult, but necessary in the cork industry. After previous studies, we know that cork stopper/disk quality can be detected by using several features obtained from the cork images: cork texture, defect area, etc. In this work we analyse the performance of a combination of image processing and a neuro-fuzzy classifier for the cork industry. As conclusion we can state this new neuro-fuzzy classification system widely improves our previous results and those obtained by other authors in related researches, being suitable for this industry because it presents a good response to classification problems with class overlapping.