<![CDATA[Katsuwonus pelamis or skipjack tuna is one of the results of fishing commodities from Gorontalo Province. The quality of fresh fish can be degraded easily if not handled and stored properly. Thus, in this study an automatic system for classifying the freshness level of skipjack tuna based on digital image processing techniques was introduced. It uses Local Binary Pattern (LBP) to extract local texture features and Gray Level Co-occurrence Matrix (GLCM) for statistical texture analysis with classification done by K-Nearest Neighbor (K-NN) algorithm using Euclidean distance as a measurement between features. There were 819 training images and 140 test images used in four categories: Fresh, Not Fresh, Worth Consuming, and Rotten. Tests on several values of k showed that the highest accuracy was at k = 1 with an accuracy rate of 86.42% while the lowest was at k = 9 with a rate of 49.28%. This indicates that the combination LBP-GLCM applied in K-NN has potentiality to capture texture difference effect from various levels fish freshness. This method is non-destructive and could be onboard application for fish quality monitoring as well as automatic system for freshness evaluation.]]>
Copyrights © 2025
