Diseño de un sistema de cuantificación automática de biomasa basado en procesamiento de imágenes y fotogrametría con vehículos aéreos no tripulados

Abstract

This work presents the design and validation of a software system able to count trees automatically, applying image segmentation and machine learning techniques to two-dimensional photogrammetric products generated by an image sensor on board an UAS system. Through a photogrammetric flight methodology, it is possible to generate a digital elevation model and orthomosaic of the region of interest. Which then, are processed by the software system to extract, identify and quantify tree physical properties in this case. Finally, the obtained properties are used for a biomass estimation according to a state of art theoretical model.