查看文章

Recognition system is an essential combat of computer science. Fruit recognition is one of them. Researcher have been roaming around this area for a decade now. Previously, many traditional machine learning techniques and deep learning methods have been employed for successful recognition of fruits and in some cases, a high accuracy have been achieved. However, in our study we showed that our proposed working diagram outperformed all the previous studies. In this paper, we started with Fruit-360 dataset of 109 distinct classes and applied three feature extraction methods (hu moments, haralick texture and color histogram). After that we applied several machine learning approaches (Decision Tree, K-Nearest Neighbors, Linear Discriminant Analysis, Logistic Regression, Naïve Bayes, Random Forest and Support Vector Machine) to train the models. Finally, the test results were calculated and K-Nearest Neighbor along with Random Forest classifiers produced best results with a false positive rate of 0%, hence, achieving a high accuracy.