Palm Oil Fresh Fruit Bunch Ripeness Grading Recognition Using Convolutional Neural Network

Abstract

This research investigates the application of Convolutional Neural Network (CNN) for palm oil Fresh Fruit Bunch (FFB) ripeness grading recognition. CNN has become the state-of-the-art technique in computer vision especially in object recognition where the recognition accuracy is very impressive. Even though there is no need for feature extraction in CNN, it requires a large amount of training data. To overcome this limitation, utilising the pre-trained CNN model with transfer learning provides the solution. Thus, this research compares CNN, pre-trained CNN model and hand-crafted feature and classifier approach for palm oil Fresh Fruit Bunch (FFB) ripeness grading recognition. The hand-crafted features are colour moments feature, Fast Retina Keypoint (FREAK) binary feature, and Histogram of Oriented Gradient (HOG) texture feature with Support Vector Machine (SVM) classifier. Images of palm oil FFB with four different levels of ripeness have been acquired, and the results indicate that with a small number of sample data, pre-trained CNN model, AlexNet, outperforms CNN and the hand-crafted feature and classifier approach.