Eventos Anais de eventos
COBEM 2019
25th International Congress of Mechanical Engineering
Machine Vision Algorithm For Coffee Fruits Count
Submission Author:
Ingrid Lorena Argote Pedraza , SP
Co-Authors:
Ingrid Lorena Argote Pedraza, Leonardo Bonacini, Alexandre Senni, John Archila, Mario Luiz Tronco
Presenter: Ingrid Lorena Argote Pedraza
doi://10.26678/ABCM.COBEM2019.COB2019-2187
Abstract
The aim of this research is the use of a vision system based in a colored monocular camera to acquire images of coffee trees and apply an algorithm known as a watershed transform to perform fruit counting on each image. To perform the image segmentation was proposed two steps based on the use of RGB and HSV color space. Prior to the use of the Watershed-transform, thresholds have been defined for each component of the RGB color space, with these thresholds it is possible to remove leaves and branches from the image, at a later stage values in the RGB space are used to segment the fruits. From the HSV space the component H (hue) is used to segment the color related to the fruits. The results of the segmentation using each pairing are combined and a filtering is performed to improve the results and thus perform a binarization. After obtaining a binary image, a watershed marked-controller transform is applied to fill the regions, and each of the particles from the transform in the image is counted as a fruit.
Keywords
Coffee Fruits, Fruit Recognition, Watershed-Transform, Machine Vision Systems
