A Systematic Literature Review of Palm Oil Maturity Detection via Deep Learning Approaches

Mohamad Ekram Nordin, Md Yushalify Misro
dx.doi.org/10.17576/apjitm-2026-1501-15

Abstract

An accurate classification of oil palm fresh fruit bunches (FFB) by maturity level is crucial for maximizing the oil yield and maintaining product quality. While manual inspection remain widely practice, it is limited by the subjectivity and inefficiency, especially in large-scale plantation environments. As a result, non-destructive approaches that incorporates computer vision and sensor technologies have emerged as promising alternatives. Hence, this review aims to (1) summarize existing studies that have applied deep learning to detect oil palm maturity, (2) examine the methodological and practical challenges that hinder the real-world deployment of these systems, and (3) discuss the future directions to advance the oil palm industry. The analysis demonstrates that despite deployment challenges, YOLO hold strong potential for automating palm oil maturity assessment. By consolidating existing knowledge and highlighting critical research gap, this review provides a clearer understanding of the field and oulines a pathway for transitioning these technologies from experimental validation to reliable field deployment.

keyword

Palm Oil; Fresh Fruit Bunches (FFB); maturity detection; YOLO; deep learning

Area

Pattern Recognition