Sugarcane Harvesting Quality by Digital Image Processing

• Autor(a) principal: De Moura Araújo, Guilherme
• Data de Publicação: 2020
• Outros Autores: dos Santos, Fernando Ferreira Lima, de Almeida, Samira Luns Hatum [UNESP], Martins, Rodrigo Nogueira, Voltarelli, Murilo Aparecido, Paixão, Carla Segatto Strini, de Assis de Carvalho Pinto, Francisco
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1007/s12355-020-00867-2; http://hdl.handle.net/11449/199220
• topic: Automation; Cut quality; Geometric properties; Modelling

Sugarcane, a major crop in the Brazilian agribusiness sector, has had significant changes in its cultivation over the years. The harvest of sugarcane was usually carried out manually. However, due to the great demand for labours and high health hazards, manual harvesting was replaced by a mechanized process. Harvesters have a basal cutting mechanism composed of blades that are responsible for cutting the sugarcane. The blades must be continuously analysed and replaced if necessary. The analysis of these blades is performed qualitatively, in which the operator analyses the current conditions of the blades (based on prior experiences); or quantitatively, using a digital calliper to measure the thickness of the blade, wherein the increase in thickness indicates greater wear. Both approaches require the harvester to be stationary and are extremely time-consuming. In order to obtain more efficient results, the present study proposes a novel method to analyse blade wear and harvesting quality through the use of digital image processing. The change in geometric characteristics (area, perimeter, rectangularity, and length) of the blades over time was evaluated along with blade wear and harvesting quality indices (damage index to stalks). The obtained results indicated that the proposed methodology was effective at assessing the quality of the harvesting operation of sugarcane and the wear in the mechanisms of basal cutting present in sugarcane harvesters. The error in estimating the damage index ranged between 0.030 and 0.033%.