Local Absolute Vertical Accuracy Computation of Wide-Coverage Digital Terrain Models

Abstract

The use of Digital Terrain Models (DTMs) as a computerized mapping and modeling infrastructure of our natural environment is rapidly and intensively growing. Many different sources of wide-coverage DTMs differ significantly in terms of geometric attributes, resolution and accuracy. These sources often describe the same coverage area by discrepant representation of the topography. Studies have shown that the inner-vertical accuracy of an individual wide-coverage DTM is not homogeneous, and cannot be expressed by a single global accuracy value. Still, in terms of quality analysis, most studies continue to investigate the accuracy of topographic models globally while ignoring inner-quality aspects. This paper suggests a method for spatial regional inner-accuracies analysis that is based on terrain characteristics and error-propagation computation via a simultaneous relative DTMs comparison mechanism. We present the evaluation of data-error characterization and local trend detection, which is an improvement in terms of DTM data quality analysis, since it addresses the vertical absolute accuracy measures on a regional scale—and not globally. The research presents the feasibility to identify regions within DTMs that show errors and data-reliability anomalies. The paper presents the outcome produced by applying this methodology, with geostatistical comparisons to the common global scheme, demonstrating the feasibility and effectiveness of the proposed algorithms.