REAL-TIME 3D MAPPING AND LOCALIZATION OF PALM OIL TREE FOR HARVEST DATA MANAGEMENT SYSTEM

Abstract

The palm oil industry, particularly vital in Asian nations such as Malaysia and Indonesia, plays a pivotal role in their economies. Its efficiency hinges on effective palm oil harvest management. This industry is poised for substantial growth, with global palm oil demand projected to surge from 51 million tons to an estimated 120 to 156 million tons in the next three decades. However, it faces significant challenges, notably its heavy reliance on foreign labor, particularly for the labor-intensive tasks of harvesting and assessing fruit bunch maturity, a situation further exacerbated by recent labor shortages due to the COVID-19 pandemic. Therefore, this study proposed a comprehensive approach to 3D mapping and tagging, centered around the detection and localization of palm oil trees. The key methodology involves employing object detection technology to extract the coordinates of these trees. Subsequently, the RTAB-Map is harnessed to precisely localize these identified palm oil trees within a 3D space. The results of this research demonstrate a robust and efficient system for palm oil tree detection and 3D localization. By utilizing object detection and RTAB-Map, we achieve high accuracy in tree identification and localization, laying the foundation for improved management of palm oil plantations. This technology not only facilitates the monitoring and assessment of palm oil tree health but also enhances overall plantation management, resource allocation, and sustainability efforts. The findings herein signify a significant step towards optimizing palm oil cultivation practices and, consequently, fostering environmental conservation and sustainable agricultural practices within the industry.