Machine Vision for Starfruit (Averrhua Carambola) Inspection

This paper describes a machine vision system for starfruit inspection. We propose a setup framework for image acquisition. The acquired image is then analyzed using convex hull algorithm to determine the shape defect and classify the starfruit. This paper proposes a hardware design to capture starfruit images from three sides so as to obtain overall shape of the fruit. The framework was simulated using Autodesk 3dsmax 2010 software and prototype hardware was built to verify the design.