<![CDATA[Oil palm tree census plays a crucial role in plantation management, influencing production planning and managerial decision-making. Manual counting using drone imagery in QGIS is time-consuming and prone to human error. This study aims to design and implement an automatic oil palm tree census detection system using the YOLOv11 algorithm integrated with QGIS. A quantitative approach with system development methodology was employed. Drone imagery was processed in Agisoft Metashape to produce orthomosaic GeoTIFF images, then annotated via Roboflow and trained using YOLOv11 for 100 epochs. Detection was performed using SAHI (Sliced Aided Hyper Inference) with 640×640 pixel tiles (25% overlap) and DBSCAN deduplication. The system successfully detected 4,886 oil palm trees in Block P14402 Sei Baleh Estate PT BSP (±24 Ha). Model evaluation yielded Precision 0.96, Recall 0.95, mAP@IoU=0.5 of 0.98, and F1-Score 0.955. Black Box Testing across 9 functional scenarios all passed. The automated system proved significantly more efficient than the manual method, supporting digitalization of oil palm plantation management.]]>
Copyrights © 2026
