A Variational Approach to Semiautomatic Generation of Digital Terrain Models

Unger, Markus; Pock, Thomas; Grabner, Markus; Klaus, Andreas; Bischof, Horst

doi:10.1007/978-3-642-10520-3_107

Cited by 14 publications

(14 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The area covered in the field survey was approximately 710 m × 210 m (length × width) along the river course, and the flight altitude was 70 m. The Pix4D photogrammetric process steps include the estimation of the UAV's camera parameters for image calibration and bundle adjustment, the extraction of identical image points (tie points) between the overlapping pictures, the estimation of the 3D point cloud and the built up of the DSM. The photogrammetry algorithms are analytically described in Unger, Pock, Grabner, Klaus, and Bischof (), and the mean error of the reconstructed surface is approximately 1–3 times the ground sampling distance (Küng et al, ). The cell size of the produced DTM was approximately 0.03 m, which corresponds to a potential inherent error between 0.03 m and 0.09 m. The number of 2D keypoints observations used by the photogrammetric algorithm (matched in at least two images) is 1,063,617, and the number of 3D points generated is 400,708.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Sensitivity of habitat hydraulic model outputs to DTM and computational mesh resolution

Papaioannou

Papadaki

2019

Ecohydrology

View full text Add to dashboard Cite

In this study, a state‐of‐the‐art approach in modelling fish habitats, using high‐resolution topographical data, obtained from unmanned aerial vehicle, was applied. Habitat Suitability Indices are used to predict how changes in discharge affect instream fish habitats. Habitat Suitability Indices regarding depth and velocity for two size classes (small sized fish 5–15 cm total length and large sized >15 cm total length) of Salmo pelagonicus and Barbus balcanicus were used, in combination with a two‐dimensional hydraulic‐hydrodynamic model, for the estimation of the weighted usable area (WUA) in a mountainous stream. Computational mesh and/or digital terrain model (DTM) resolution selection may influence the accuracy of WUA results, especially in boulder and cobble‐bed streams with complex habitat structures. The aim of the study is to examine the sensitivity of various hydraulic‐hydrodynamic modelling geometry configurations on WUA at ungauged or poorly gauged streams. Comparisons of three different geometry configurations: (1) identical computational mesh and DTM resolution (SensComb); (2) finest computational mesh resolution combined with different DTM resolutions (SensDTM); (3) finest DTM resolution combined with different computational mesh resolutions, as part of two‐dimensional hydrodynamic modelling, were applied to test the differences in WUA (SensMesh). WUA maps were generated for both fish species and class sizes for each modelling geometry configuration and compared with each other for assessing the sensitivity of the two‐input data (computational mesh and DTM). Results provided by both indices and their spatial distribution indicated the optimal DTM and computational mesh resolution as well as the sensitivity of a specific hydraulic‐habitat model on them.

show abstract

Section: Methodsmentioning

confidence: 99%

“…The field topographic data in the study area (Drosopigi river) were col- Unger, Pock, Grabner, Klaus, and Bischof (2009), and the mean error of the reconstructed surface is approximately 1-3 times the ground sampling distance (Küng et al, 2012).…”

Section: Field Data Collectionmentioning

confidence: 99%

Sensitivity of habitat hydraulic model outputs to DTM and computational mesh resolution

Papaioannou

Papadaki

2019

Ecohydrology

View full text Add to dashboard Cite

show abstract

“…A classificação e filtragem automática da nuvem de pontos baseiam-se no ângulo, altura e distância de cada ponto para diferenciar o relevo das demais feições (vegetação e construções) (UNGER et al, 2009;WESTOBY et al, 2012;TURNER et al, 2012;PIX4D, 2017) HUNG,.M.N.W.B., SAMPAIO,T.V.M., SCHULTZ,G.B., SIEFERT,C.A.C., LANGE,D.R., MARANGON, F Entretanto, a classificação e filtragem automática pode gerar ruídos devido a densidade da vegetação presente na área de estudo, fazendo com que o Pix4D Desktop classifique em alguns casos, árvores de menor altura, como um ponto de solo em áreas de mata nativa e reflorestamento de pinus, onde a vegetação apresenta diferentes alturas e assim, enfatizando as dificuldades de gerar um MDT em áreas de floresta densa. Deste modo, foi necessário realizar a filtragem dos ruídos manualmente, a partir da exclusão de pontos que fugiram da tendência da superfície do relevo.…”

Section: Materiais E Métodosunclassified

Levantamento Com Veículo Aéreo Não Tripulado Para Geração De Modelo Digital Do Terreno Em Bacia Experimental Com Vegetação Florestal Esparsa

et al. 2018

View full text Add to dashboard Cite

, Irani dos Santos 1g RESUMO A representação do relevo por meio dos modelos digitais do terreno (MDT) demanda do conhecimento das potencialidades e limitações das técnicas e instrumentos da aerofotogrametria. O levantamento com veículos aéreos não tripulados (VANT) e câmeras não métricas tem sido amplamente utilizado como método para obtenção de MDT devido a sua facilidade de operação e baixo custo. Entretanto, a geração do MDT a partir do modelo digital de superfície exige cuidados e pode adicionar erros ao modelo. Neste contexto, este artigo tem por objetivo avaliar a qualidade do MDT gerado por aplicação de procedimento semiautomático sem pontos de controle em campo, em área com presença de vegetação esparsa. O modelo foi gerado com o VANT Phantom 3 Professional com altura média de voo de 60 m e sem a utilização de pontos de controle. Para a avaliação da acurácia e validação do MDT foram utilizados 354 pontos obtidos por técnicas de levantamento topográfico convencional. As diferenças altimétricas entre o MDT e o levantamento foram, em geral, inferiores a 0,5 m (R² = 0,99), sendo obtido um erro padrão de 0,29 m e um padrão de exatidão cartográfica de 0,49 m, equivalente a um levantamento altimétrico classe A (1:5.000). Foi ainda elaborado um modelo digital da vegetação, sendo validado a partir de dados obtidos em campo para 54 árvores (R²= 0,87 e erro padrão = 0,54 m). Os resultados evidenciam a aplicabilidade da utilização de levantamento com VANT para obtenção de MDT em locais com vegetação florestal esparsa mediante a validação dos resultados com levantamento topográfico convencional.Palavras-chave: modelagem digital do relevo, imagens de alta resolução, aerofotogrametria. ABSTRACTTerrain representation through digital terrain models (DTM) requires the knowledge of the potentialities and limitations of aerial photogrammetry techniques and devices. Unmanned aerial photogrammetric surveys using photos acquired by non-metric cameras has been widely used as a method to generate DTMs due to its usability and low cost. However, some wariness for DTM generation from the digital surface model (DSM) resulting from aerial photogrammetric surveys should be taken due the errors that could be added on this procedure. Here, the main objective was evaluate the DTM generated with a semi-automatic procedure without ground control points for a sparse forested area. The survey was carried out using an unmanned aerial vehicle (UAV) model Phantom 3 Professional with average flight height of 60 m and without ground control points. 354 elevation points acquired by a topographical survey were used for cartographic accuracy assessment and DTM validation (spatial resolution = 63 cm). The height differences between DTM and topographical survey elevation points were lower than 0,5 m with a standard error (SE) of 0,29 m and a standard cartographic accuracy of 0,49 m, corresponding to A class altimetric data survey (1:5.000). Digital forest canopy height model was derived from DTM and DSM and validated through field survey of 54 trees ...

show abstract

“…The main difference to spaceborne data is that the underlying DSMs are of significantly higher resolution and of higher accuracy such that an image classification also works very robust -which would not be the case for spaceborne data. (Unger et al, 2009) present a variational approach which uses a DSM as single input. The basic concept is to extract a very smooth surface using a strong regularization weight within a variational formulation.…”

Section: Airborne Lidar Datamentioning

confidence: 99%

Advanced DTM Generation From Very High Resolution Satellite Stereo Images

Perko

Raggam

Gutjahr

et al. 2015

ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:This work proposes a simple filtering approach that can be applied to digital surface models in order to extract digital terrain models. The method focusses on robustness and computational efficiency and is in particular tailored to filter DSMs that are extracted from satellite stereo images. It represents an evolution of an existing DTM generation method and includes distinct advancement through the integration of multi-directional processing as well as slope dependent filtering, thus denoted "MSD filtering". The DTM generation workflow is fully automatic and requires no user interaction. Exemplary results are presented for a DSM generated from a Pléiades tri-stereo image data set. Qualitative and quantitative evaluations with respect to highly accurate reference LiDAR data confirm the effectiveness of the proposed algorithm.

show abstract

A Variational Approach to Semiautomatic Generation of Digital Terrain Models

Cited by 14 publications

References 16 publications

Sensitivity of habitat hydraulic model outputs to DTM and computational mesh resolution

Sensitivity of habitat hydraulic model outputs to DTM and computational mesh resolution

Levantamento Com Veículo Aéreo Não Tripulado Para Geração De Modelo Digital Do Terreno Em Bacia Experimental Com Vegetação Florestal Esparsa

Advanced DTM Generation From Very High Resolution Satellite Stereo Images

Contact Info

Product

Resources

About