The use of elimination method and nearest neighbor for oil palm fruit ripeness indicator

Abstract

Fruit ripeness identification is hard to measure especially when it involves color as main indicator. Suitable color model must be chosen to determine the right color for the ripeness identification. Hue, Saturation and Value (HSV) are proved to be a better choice because it can define the color intensity. Besides, it also helps to choose colors which are similar to the eyes. Manual grading process by human graders at oil palm mills lead to misconduct and mistakenly claimed unripe fruits as the ripe ones. This will cause trouble when the error report arrives at the production site for the oil production sterilization process. Furthermore, research done by the Federal Land Development Authority (FELDA) in Malaysia stated that an approximation of 60% palm oil are coming from ripe fruit meanwhile 40% are from underripe and 20% from unripe fruit minus water and dirt. This is proved the importance of identifying the right fruits for the purpose of oil palm production is extremely important. This paper studies the use of elimination method and nearest neighbor for oil palm fruit ripeness indicator. Result shows value gives the best indicator by providing the highest recognition rate towards ripe and unripe category. © 2011 Springer-Verlag.