Digital Image Processing Digital Image Processing To Identify Quality Of Oil Palm Fruit Using Learning Vector Quantization Method (Case Study: Pt Jaya Gemilang Sukses)

Authors

  • Linda Linda Politeknik Caltex Riau

Abstract

PT Jaya Gemilang Sukses (JGS) is one of the companies engaged in the oil sector by producing Crude Palm Oil (CPO). In the process of processing oil palm fruit into CPO, PT JGS conducts the process of sorting palm fruit manually and subjectively by human vision. External factors such as fatigue, laziness, and limited number of workers can affect the process of sorting oil palm fruit. From these problems, a digital image processing application was built that was able to identify the quality of oil palm fruit. The digital image processing applied to the application is HSV color feature extraction (Hue, Saturation, Value) and GLCM (Gray Level Co-Occurrence Matrix) texture feature extraction. The results of the feature extraction are then input into the Learning Vector Quantization (LVQ) method, so the output of this system is the quality and maturity level of the oil palm fruit. Based on the results of the confusion matrix test, it can be concluded that the extraction of GLCM texture features is not appropriate because it is susceptible to rotation and cause the accuracy not too high at 43.75% in the maturity level classification and 78.125% in the quality classification. Based on usability testing, a percentage of 81.33% was obtained, can be concluded that workers at PT JGS strongly agreed that the system helped in the process of sorting fruit at PT JGS. Based on testing the LVQ method calculation, it can be ascertained that the source code used is correct and according to the algorithm.

Published

2020-08-28

Issue

Section

Artikel