Real time visual system for starfruit maturity index classification

Abstract

Considering the demand for fresh starfruit becomes higher nowadays, an automatic system for starfruit quality inspection is mostly needed since the quality inspection of starfruit is still manually performed by human. The objectives of this thesis are to develop a real time visual system and to implement the colour maturity algorithm into a Field Programmable Gates Array (FPGA) system for starfruit colour maturity classification. Generally, the system designed in this work consists of three main sub-systems: input, process and output. The input of the system is acquired by using a digital camera with YCbCr format. The second part of the system is the main processing system which is FPGA. The processes on the FPGA can be divided into three parts, which are segmentation, feature extraction and classification. The segmentation process is utilised to determine the Region Of Interest (ROI) of the starfruit area by using fixed threshold value based on Cb component. The feature fed to the system is extracted from Cr component that becomes the input to the proposed rule-based classifier. In this work, the starfruit is classified into 6 maturity levels. The performance of the proposed system achieved 89% of starfruit correctly classified.