Smart farming takes place by using recent UAV-technology for agricultural profitability. Multiple companies and research centers are working on realizing real scenarious. But there are only restricted experiences about how to process the data faster and better in real-time. Autonomous systems that can decide instanteously after receiving sensor information could increase the farming efficiency. The agroindustry research group from the Escuela Colombiana de Ingeniería and Universidad de Ciencias Aplicadas y Ambientales en Bogota-Colombia in alliance with the University of Applied Sciences Wuerzburg-Schweinfurt are committed to that idea. This project uses an NVIDIA Jetson embedded board to process agricultural information in order to find abnormalities. In this article we prove that this board is able to make the calculations for a vegetation index in real-time from multi-spectral images.