Sistema eficiente y de bajo costo para la selección de granos de café: una aplicación de la visión artificial

• Main Authors: Cesar Wilfredo Rosas-Echevarría, Hubel Solís-Bonifacio, Alberto Franco Cerna- Cueva
• Format: Article
• Language: English
• Published: Universidad Nacional de Trujillo 2019-09-01
• Series: Scientia Agropecuaria
• Subjects: coffee beans; computational vision; grain selection; low cost.
• Online Access: http://revistas.unitru.edu.pe/index.php/scientiaagrop/article/view/2554

Quality is an important factor when positioning a product in the market, so it is necessary to have an efficient selection process. Currently there are expensive equipment that carry out the selection process that is not accessible to small businesses, it is therefore mentioned that a low-cost procedure is proposed through which the coffee bean selection process can be performed, using a library Real- time image processing for quality selection based on color and size (pixels on screen). A sample of 50kg of coffee was used for manual quality selection based on the proportion of colors (green, red, brown colors), size per grain that is classified into three categories (< 1 cm2 , = 1 cm2 , > 1 cm2 ). Then the same process was carried out on the same sample by using the image processing system based on color and size (pixels per screen by rows and columns for the calculation of the area of each grain) by a belt where 3 grains passed per second. The results referring to the color-based quality did not show a significant difference α = 0.05, but with respect to the time used, the selection process through image processing lasted one hour compared to the 2 hours of the manual process, therefore the selection of the Coffee beans using the image processing system is superior to the manual selection process.