Recognition of ripe, unripe and scaled condition of orange citrus based on decision tree classification

Orange, which is most cultivated fruit in the world, is commonly used in food processing industries to prepare juice, marmalades, and orange pulp. With modern computer vision techniques, manual sorting of fruits is being replaced with automated low cost and consistent approach. This paper presents a mean for distinguishing orange condition (ripe, unripe and scaled or rotten) rapidly. Fruit image features including RGB color space and gray values based on BIC (Border/Interior pixel Classification) are extracted. An investigation for the applicability and performance of various classification algorithms including Naïve Bayes, Artificial Neural Network, and Decision Tree has been performed. Comparisons among results of these algorithms have been drawn and it has been observed that Decision Tree classification technique for orange conditions is efficient than other techniques. The results recorded for the accuracy, precision, and sensitivity using this technique are 93.13%, 93.45%, and 93.24% respectively.