Serviços Personalizados
Journal
Artigo
Indicadores
Citado por SciELO
Acessos
Links relacionados
Similares em SciELO
Compartilhar
Revista cartográfica
versão On-line ISSN 2663-3981versão impressa ISSN 0080-2085
Resumo
RODRIGUEZ AVI, José. Characterization DEM Error by Distribution mixtures. Rev. cartogr. [online]. 2021, n.103, pp.123-143. Epub 27-Set-2021. ISSN 2663-3981. https://doi.org/10.35424/rcarto.i103.968.
The positional accuracy of Digital Elevation Models (DEM) is usually assessed by analyzing the vertical error observed in point samples. The traditional statistical techniques suppose that altimetry errors verify normal distributions, but, in practice, this is not true in many cases. In this work, the use of techniques based on the determinacy of Gaussian finite mixture models is presented in a novel way in the field of DEM, to better approximate the observed error distribution. A methodology to apply the process is proposed ad it is applied to real data from DEM of 2x2 m spatial resolution (reference) and 5x5 m (product).
Palavras-chave : DEM; spatial accuracy; Finite mixture models; normal distribution.