Evaluation and Detection of Gaps in Curved Sugarcane Planting Lines in Aerial Images

Sugarcane is one of the main crops in the world due to the economic value it promotes by selling its derivatives. A diversity of technologies has been developed to optimize agricultural activities and maximize the productivity of sugarcane crops. In this sense, our primary goal is to contribute to this research area by detecting planting lines and measuring their faults, including the evaluation of curved lines that substantially limit numerous solutions in practical applications. An automatic method that identifies and measures sugarcane planting lines through digital image processing techniques and machine learning algorithms is presented. The proposal is evaluated using a database of real scene images, which were classified by K-Nearest Neighbors (KNN) and prepared with the support of a small unmanned aerial vehicle (UAV). Experimental tests show a low relative error of approximately 1.65% compared to manual mapping in the planting regions. It means that our proposal can identify and measure planting lines accurately, which enables automated inspections with high precision measurements.