Convolutional neural network for optimal pineapple harvesting

• Main Authors: Azman, Ahmad Aizuddin, Ismail, Fatimah Sham
• Format: Article
• Language: English
• Published: Penerbit UTM Press 2017
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utm.my/id/eprint/80309/1/FatimahShamIsmail2017_ConvolutionalNeuralNetworkforOptimalPineapple.pdf; http://eprints.utm.my/id/eprint/80309/; https://elektrika.utm.my/index.php/ELEKTRIKA_Journal/article/view/54
• Institution: Universiti Teknologi Malaysia

Upon ripening, colour of pineapple’s peel gradually changes from green to yellowish, which spreading from bottom to the top. The objective of this project is to develop a computational intelligence method for pineapple maturity indices classification for optimal harvasting. Pineapple maturity indices can be grouped into three levels, which are unripe, partially ripe and fully ripe for determining optimal pineapple harvesting. Previous works on classifying fruit’s ripeness rely on manual hand-engineered feature extraction and selection. This project proposes new intelligent method using convolutional neural network (CNN) that has the ability to learn several unique features from the given task automatically through supervised learning. The simulation results show that the method achieved 100% classification’s accuracy for determining unripe and fully ripe level and 82% accuracy for partially ripe level.