<![CDATA[Oil palm plantations require accurate and timely inventory data to support plantation management, productivity assessment, and sustainable agricultural practices. However, manual tree inventory in large plantation areas is labor-intensive, time-consuming, and prone to human error. This study proposes an automated approach for oil palm crown detection and tree counting using high-resolution Unmanned Aerial Vehicle (UAV) imagery. To improve image quality, Contrast Limited Adaptive Histogram Equalization (CLAHE) was applied as a preprocessing step, and the Roboflow Detection Transformer (RF-DETR) was used as the object detection model. The proposed method was evaluated using 1,135 UAV images containing 56,547 annotated oil palm crowns collected from commercial plantations in West Kalimantan, Indonesia. Experimental results demonstrate that the proposed approach achieved mAP@50 of 97.5%, precision of 97.1%, recall of 96.0%, and F1-score of 96.6% for oil palm crown detection. In the tree-counting evaluation, the system successfully detected 1,105 of 1,120 ground-truth trees, achieving an overall accuracy of 98.5%. Furthermore, the proposed method achieved an average detection time of 16.1 ms, indicating high computational efficiency. These results demonstrate that the proposed framework provides an effective and practical solution for automated oil palm inventory and plantation monitoring using high-resolution UAV imagery. Contribution to Sustainable Development Goals (SDGs): SDG 2: Zero Hunger SDG 9: Industry, Innovation and Infrastructure SDG 12: Responsible Consumption and Production SDG 15: Life on Land]]>
Copyrights © 2026
