Using Machine Vision To Inspect Oil Palm And White Powder Starch

Color and size are two important attributes for quality evaluation of most food and bioproducts. Many of the quality factors affecting oil palm and white powder starch are determined primarily by visual inspection. Such inspections determined market price and to a certain extent consumers' preferences. However, the inspection of oil palm and white powder starch is a very repetitive and tedious task which is also subjective in nature. This paper describe techniques and methods developed for starch and oil palm imaging using machine vision technology. The system was able to correctly classify 91.2% of oil palm for color, and 91.1% of starch for size. Classification scores located near the separation boundaries were the major source of misclassification error. Key-Words: Machine vision, Automated inspection, Color analysis, Stepwise discriminant, Quality evaluation.