Machine learning based quality evaluation of mono-colored apples

Abstract

In the field of agriculture science, automatic visual inspection improves the commercial, quality and fertility of the country. It is very challenging to sort the fruit based on quality because of varieties of fruits available in the market. Human grades the fruit but it is inconsistent, stagnant, and expensive and influenced by the surrounding. Thus an effective system for grading of fruit is desired. In this paper, an automated fruit grading system is developed for apple to classify based on external quality. The different combination of several features are considered depending on the damages exposed on apple fruits. In this work, these features are considered as input to train Support Vector Machine (SVM). The classifier has been contemplated with two different database of apple: one having 100 color images out of which 24 are of apples with various defects and the other dataset having 112 color images out of which 56 are of apples with various defects. The system performance has been validated using k-fold cross validation technique by considering different values of k. The maximum accuracy 96.81% and 93.00% for two dataset respectively, achieved by the system is encouraging and is comparable with the state of art techniques.