Answers are grounded in this paper and cite their sources.
As the demand for food and renewable energy grows, the need for more efficient and sustainable agricultural practices becomes urgent. Sugarcane, a significant crop for both food and biofuel production, presents challenges during mechanised harvesting due to its long growth cycle and sensitivity to mechanical damage. Current GNSS-based autonomous harvesting systems often lack the precision needed, leading to crop losses from stump crushing and soil compaction, directly affecting yield and sustainability. This paper proposes a UAV-assisted image-based method to estimate sugarcane row orientation, improving autonomous harvester guidance accurately. A labelled aerial image dataset was built through expert manual annotations, probabilistically fused using a Static Kalman Filter to generate a reliable ground truth. The image processing pipeline integrates the Excess Green Vegetation Index (ExG), morphological filtering, and a rotation-based vertical projection, refined with a Savitzky-Golay filter, to detect crop row angles robustly. The system was implemented on an embedded processor (ARMv8) platform, demonstrating field-ready performance. Experiments under diverse field conditions achieved a median angular error of 0.9°, average processing time of 2.82 seconds per image, and energy consumption of only 5.97 W, meeting the operational tolerances for UAV-assisted orientation correction. This study presents a validated heuristic framework, a robust expert-fused dataset, and a real-time embedded solution, demonstrating the potential of UAV-assisted systems to enhance harvesting efficiency, mitigate losses, and support sustainable precision agriculture.
No takes yet. Share an insight, caveat, or question.
Ferreira et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: