AI-Driven Quality Control and Ripeness Classification in the Palm Oil Industry: A Systematic Review of Computer Vision and Deep Learning Methods

This research is motivated by the need to improve the quality and consistency in assessing the ripeness of fresh fruit bunches (FFB) of oil palm, which has so far relied on subjective manual inspection. This study aims to synthesize the development of artificial intelligence-based methods, particularly computer vision and deep learning, in ripeness classification and quality control. The methodology used is a PRISMA-based systematic literature review of 37 scientific articles from the period 2020–2026. The results show a significant shift from conventional machine learning methods to deep learning models, especially CNN, with accuracy generally above 90%. The performance of the model is greatly influenced by the quality of the dataset, validation, and implementation conditions. Additionally, the application of real-time systems based on mobile and conveyor technology is beginning to support industry integration. The implication is that AI adoption has the potential to enhance efficiency, consistency, and transparency in the quality management of the oil palm sector, although it still requires data standardization and cross-location validation.