Services on Demand
Journal
Article
Indicators
- Cited by SciELO
- Access statistics
Related links
- Cited by Google
- Similars in SciELO
- Similars in Google
Share
Prospectiva
Print version ISSN 1692-8261
Abstract
HERRERA PEREZ, Jean Carlos et al. Classification of coffee fruits based on ripeness and broca detection using image processing techniques. Prospect. [online]. 2016, vol.14, n.1, pp.15-22. ISSN 1692-8261. https://doi.org/10.15665/rp.v14i1.640.
In the present article is proposed the development of two algorithms, one to determine if a coffee fruit is proper for production based on its color (ripeness) and the other one to detect the presence of the "broca" plague. The color classifier consists of several stages: a knowledge database, which has a bank of coffee-fruits images in ripe and unripe state; a stage of pre-processing to clean up noise in the image; next, the segmentation process to extract the object of study, then follows the color-characteristics extraction process. Finally, the classifier, which consists of an artificial neural network where fruits are classified ripe or unripe. On the other hand, the broca detection algorithm was developed by means of binarization, in order to search for dark zones in the image, as it is the aperture made by the broca on the coffee. The artificial neural network proposed had an effectiveness of 97% at detecting the ripe state of the coffee fruits, therefore demonstrating the viability and the minimal invasion of the method proposed for the quality control of coffee fruits.
Keywords : Feature Extraction; Artificial Neural Networks; Image color analysis; Image analysis; Image segmentation; Image classification; Machine vision.