Fuzzy Classification of Color Carrots (Dacus Carota) using Raspberry Pi towards Farming 4.0

• Published in: 2023 IEEE 13th Annual Computing and Communication Workshop and Conference (CCWC) pp. 1174 - 1178
• Main Authors: Villasenor-Aguilar, Marcos J., Padilla-Medina, Jose A., Prado-Olivarez, Juan, Martinez-Diaz, Saul, Mendez-Gurrola, Iris-Iddaly, Barranco-Gutierrez, Alejandro I.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 08.03.2023
• Subjects: Agriculture; Carrots; Energy consumption; Fuzzy Logic; Image color analysis; Pixel; Raspberry Pi 3; Sensors; Training; Wireless communication; Wireless sensor networks
• DOI: 10.1109/CCWC57344.2023.10099149

Precision agriculture is the application of data analytics and communication technology in agriculture, with the purpose of obtaining information from the different sensors in the field, and this evidence can be consulted by the user using mobile devices. In that respect, the determination of the quality of carrots in harvest requires the use of portable electronic systems with low energy consumption and scalable to carry out the traceability of logistics processes. The objective of this work is to automatically classify color carrots, using color images processed with a Raspberry Pi 3 programmed in Python using Fuzzy Logic. The methodology used employed four phases, the first phase consisted of obtaining the modeling data using carrot of white, yellow, purple and orange colors. The second phase is responsible for processing the images of the modeling set to obtain the descriptors used in the Fuzzy model. Subsequently, the Fuzzy classification model was designed using the descriptors associated with the purple, green, yellow and orange tones; The last phase was evaluation of the precision agriculture system. This correctly classified the modeling set consisting of 60 samples in the different shades, of which 45 samples were used for training the Fuzzy system and the rest for validation. It was found that the use of the ratio of areas by color allows the variety of samples to be correctly classified. In addition, it was observed that the Raspberry Pi development platform can handle sensor architecture in an integrated way and the use of the IoT Thing Speak platform that facilitates the monitoring of sample classification. The use of the Raspberry Pi and the Thing Speak platform facilitates the construction of the Farming 4.0 system that allows automating the quality process.