A Computational Comparative Analysis Between Nvidia Jetson Nano and Raspberry Pi CM4 for the Classification of White Asparagus with SVM

• Autores: Ruiz, Edgar, Ortiz, Manuel, Vinces, Leonardo
• Formato: artículo
• Fecha de Publicación: 2022
• Institución: Universidad Peruana de Ciencias Aplicadas
• Repositorio: UPC-Institucional
• Lenguaje: inglés
• OAI Identifier: oai:repositorioacademico.upc.edu.pe:10757/660554
• Enlace del recurso: http://hdl.handle.net/10757/660554
• Nivel de acceso: acceso embargado
• Materia: Asparagus; Image processing; Jetson Nano; Raspberry Pi CM4; SBC; Support Vector Machine

The following study proposes to analyze and compare the computational response provided by the Single Board Computers (SBC) Raspberry Pi CM4 and Nvidia Jetson Nano because those are important elements in machine learning applications and implementation of automated systems. Both were chosen due to their similar specifications to achieve a fair comparison. For the development of this research, an algorithm was implemented with a Support Vector Machine to be able to compare the performance in real-time of both computers based on performance metrics such as execution time, algorithm accuracy, CPU performance, and temperature. To validate results, there is a database of 2186 white asparagus images, which were classified based on attributes such as length, curvature, diameter, and purple hue. These attributes are established by the Peruvian Asparagus and Vegetable Institute (IPEH) in the Peruvian Technical Standard to ensure the quality of fresh asparagus for export. The algorithm has been designed to classify asparagus according to this technical standard.