Sistema de visión artificial para la identificación del estado de madurez de frutas (granadilla)
Fecha
Autor corporativo
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Distrital Francisco José de Caldas
Compartir
Director
Altmetric
Resumen
Descripción
El manejo adecuado de frutas se ha convertido en una de las actividades económicas más importantes en la agricultura colombiana [1]. A la fecha, la identificación del estado de maduración de frutas se realiza manualmente [2], presentando variabilidad por la subjetividad producida debido a la fatiga ocular del experto. El propósito de esta investigación fue desarrollar una herramienta computacional para la identificación del estado de maduración de granadillas a partir del reconocimiento de imágenes. El área en píxeles de las imágenes perteneciente a la fruta fue extraída mediante la técnica de Otsu usando librerías de OpenCv en Python. Finalmente, la tarea de clasificación se realizó a través del análisis de agrupamiento, en el cual fueron asignados 110 puntos RGB pertenecientes a cada estado de maduración de la granadilla. Los resultados obtenidos muestran 92,6% de aciertos en la identificación del estado de maduración, a partir de un conjunto de 90 imágenes obtenidas de 90 frutas en diferentes estados de maduración, en comparación con el análisis manual acorde a lo establecido por la Norma técnica colombiana NTC 4101.
The proper handling of fruits has become one of the most important economic activities in the Colombian agriculture [1]. Actually, the identification of the ripeness of fruit is made manually [2], which induces variability due to subjectivity by expert eye strain. The purpose of this research was to develop a computational tool for identifying the state of ripeness of passion fruit (granadilla) through images recognition. The area in pixels of the fruit images was extracted by a technique called Otsu, using OpenCv libraries in Python. Finally, the task of classification was conducted through cluster analysis, here were assigned 110 points RGB belonging to each state of maturity of passion fruit. The results showed 92, 6% of accuracy for identifying the state of ripeness, from a set of 90 images obtained from 90 fruits in different stages of maturity, which was compared with traditional analysis (conducted by experts) according to the provisions of the Colombian Technical Standard NTC 4101.
The proper handling of fruits has become one of the most important economic activities in the Colombian agriculture [1]. Actually, the identification of the ripeness of fruit is made manually [2], which induces variability due to subjectivity by expert eye strain. The purpose of this research was to develop a computational tool for identifying the state of ripeness of passion fruit (granadilla) through images recognition. The area in pixels of the fruit images was extracted by a technique called Otsu, using OpenCv libraries in Python. Finally, the task of classification was conducted through cluster analysis, here were assigned 110 points RGB belonging to each state of maturity of passion fruit. The results showed 92, 6% of accuracy for identifying the state of ripeness, from a set of 90 images obtained from 90 fruits in different stages of maturity, which was compared with traditional analysis (conducted by experts) according to the provisions of the Colombian Technical Standard NTC 4101.
Palabras clave
application of artificial vision, colombian agriculture, fruit, image processing, Otsu segmentation, Python, agricultura colombiana, aplicación de visión artificial, frutas, procesamiento de imágenes, Python