TY - JOUR
T1 - Determinación de la madurez de mazorcas de Cacao, haciendo uso de redes neuronales convolucionales en un sistema embebido
AU - Heredia-Gómez, Juan F.
AU - Rueda-Gómez, Juan P.
AU - Talero-Sarmiento, Leonardo H.
AU - Ramírez-Acuña, Juan S.
AU - Coronado-Silva, Roberto A.
N1 - Publisher Copyright:
© 2020 Universidad Autonoma de Bucaramanga.
PY - 2020/7
Y1 - 2020/7
N2 - A correct cocoa harvest involves determining a pod maturity. However, this farm activity is usually handmade, using criteria such as Size and Color of the pod; those characteristics differ according to the cocoa variety, making it difficult to standardize. For this reason, this work proposes an automated method to simplify the number of variables to develop a portable, low-cost, and custom-made tool, which makes use of a convolutional neural network to indicate whether a cocoa pod is found it at the right time to harvest. The main results of this work are: 1) the construction of three labeled data sets (1992 images each), and 2) we developed an embedded system with a 34.83% mAP (mean Average Precision) accuracy. Finally, variance analysis demonstrates that image size (i.e., 4033x4033 p, 1009x1009 p, and 505x505 p) does not affect accuracy.
AB - A correct cocoa harvest involves determining a pod maturity. However, this farm activity is usually handmade, using criteria such as Size and Color of the pod; those characteristics differ according to the cocoa variety, making it difficult to standardize. For this reason, this work proposes an automated method to simplify the number of variables to develop a portable, low-cost, and custom-made tool, which makes use of a convolutional neural network to indicate whether a cocoa pod is found it at the right time to harvest. The main results of this work are: 1) the construction of three labeled data sets (1992 images each), and 2) we developed an embedded system with a 34.83% mAP (mean Average Precision) accuracy. Finally, variance analysis demonstrates that image size (i.e., 4033x4033 p, 1009x1009 p, and 505x505 p) does not affect accuracy.
KW - Cocoa
KW - Image classification
KW - Image recognition
KW - Object detection
KW - Raspberry pi
KW - Ripeness
KW - YOLO
UR - http://www.scopus.com/inward/record.url?scp=85098076195&partnerID=8YFLogxK
U2 - 10.29375/25392115.4030
DO - 10.29375/25392115.4030
M3 - Artículo Científico
AN - SCOPUS:85098076195
SN - 1657-2831
VL - 21
SP - 42
EP - 55
JO - Revista Colombiana de Computación
JF - Revista Colombiana de Computación
IS - 2
ER -