Feature Extraction and Naïve Bayes Algorithm for Defect Classification of Manalagi Apples

Abstract

Apple is one of the trees that is widely cultivated and grows in subtropical areas. In Indonesia, there are many areas that cultivate apples, including Malang, Batu, Nongkojajar. One way to increase the economic value of apple farmers is by sorting them before sending them to the market. This is important to do in order to make it easier to determine the quality and selling price of manalagi apples. Most apple sellers will sort the apples manually which results in high costs, difficulty, and inconsistency in the sorting process. So far, the classification of defects in apples has been done using the naked eye. This also requires expertise or experts in distinguishing which apple defects are. However, experts have limitations, not all apple defects can be recognized or classified. In addition, each researcher only uses one image feature, namely the texture feature. In this study, using an image dataset of Manalagi apples totaling 337 images, where there are 184 images of good apples and 153 images of defective apples by extracting features on apples, it can be concluded that the nave Bayes method can be used to classify defects of manalagi apples based on texture.