Intelligent Color Vision System For Ripeness Classification Of Oil Palm Fresh Fruit Bunch

• Main Author: Fadilah, Norasyikin
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: TK1-9971 Electrical engineering. Electronics. Nuclear engineering
• Online Access: http://eprints.usm.my/61135/1/24%20Pages%20from%2000001785141.pdf; http://eprints.usm.my/61135/
• Institution: Universiti Sains Malaysia

Ripeness classification of oil palm fresh fruit bunches (FFBs) during harvesting is important to ensure that they are harvested at the optimum stage for maximum oil production. Current harvesting methods based on observing the number of loose fruits on ground and the color of the fruits using human vision lead to subjective evaluation, laborious work, and low quality oil. Therefore, this research focuses on the development of an automated system with the ability to process the image of oil palm FFB and determine its ripeness category. The system consists of an image acquisition system, image processing component and oil palm FFB classification system. Images of oil palm FFBs of type DxP Yangambi are acquired using an IP camera which is attached to the end of a pole and connected to a computer via the RJ45 cable. The images are collected and analyzed using digital image processing techniques. k-means clustering algorithm is used to segment the image into two separate regions which are fruit and spike regions. Then, the color features of the fruit region are extracted from the images and used as inputs to an Artificial Neural Network (ANN) model learning algorithm.