Clasificación de una nube de puntos de alta densidad para la identificación de vegetación mediante un enfoque integrado

Abstract

This paper presents the validation of a point cloud generated by an aerial topographic survey conducted with an unmanned aerial vehicle (UAV), initially generating a high- resolution ortho-mosaic and georeferenced for visual validation of classification and providing a detailed geospatial context of the study area. The research focuses on the comparison of information classification methodologies in areas with different levels of vegetation density: low, medium and high. As a reference, a manual classification of the elements present in the point cloud was used to evaluate the accuracy and consistency of the results obtained through ten iterations based on deep learning methods and heuristic rules. The results include a point cloud classified according to the characteristics of the study area, detailed statistical analysis and the integration of the ortho mosaic as a validation and documentation tool, which are described in the methodological section of the document.