<![CDATA[Identification of oil palm fresh fruit bunches (FFB) based on variety is a crucial step in sorting and grading FFBs to produce good-quality crude palm oil (CPO). Most palm oil mills receive two varieties of FFBs at the reception stations, Tenera and Dura, and only a certain percentage of the Dura variety is allowed in a transporting truck. The conventional identification is destructive, cutting several fruits off an FFB bunch and checking for fruit Mesocarp and shell thickness. The method suffers a high increase in free fatty acid (FFA) content. This study is a preliminary study using computer vision and image processing to differentiate the two varieties based on their physical properties. The samples consisted of 20 Dura and 20 Tenera FFBs, 10 unripe and 10 ripe FFBs. The FFB images were acquired for both front and back sides using a color CMOS camera. ImageJ software was used to obtain the number of outer fruits and bunch surface area, used to calculate fruitlet density. Both varieties are also compared based on mass and by red, green, and blue (RGB) intensities. The results were compared to the results measured manually. The results showed that the Tenera variety exhibited higher fruit density, fruitlet count, RGB intensity compared to the Dura variety. Both varieties have higher correlations between fruit density and their masses. These results show the potential of computer vision and image processing methods to differentiate Tenera and Dura varieties, used for sorting and grading oil palm FFBs.]]>
Copyrights © 2025
