Positional Accuracy Assessment of Digital Elevation Models and 3D Vector Datasets Using Check-Surfaces

Abstract

This study focuses on the positional accuracy of Digital Elevation Models (DEMs) and 3D vector features by considering that both datasets can be used as a product to assess or as a reference. The main objective is to provide an alternative method to the traditional use of checkpoints by using check-surfaces in order to avoid identification issues. The methodology includes the determination of a set of polygons with a significant height in relation to the surrounding area (elevated or depressed) and those cells extracted from the DEM that match these elements. The check-surfaces are obtained after a triangulation of these polygons. The methodology uses procedures based on buffers to provide several results in the form of distribution functions of accuracies (2D, vertical and 3D). The trial has been carried out using a large set of data representing buildings obtained from official institutions. The results show consistent 2D, vertical and 3D accuracy values related to commonly used confidence levels. The application has demonstrated the viability of this approach for obtaining horizontal and vertical accuracies individually and jointly at any confidence level. In addition, the study includes the analysis of the results of specific zones, considering several characteristics.