Impact of Slope, Aspect, and Habitat-Type on LiDAR-Derived Digital Terrain Models in a Near Natural, Heterogeneous Temperate Forest

Original Article

Abstract

Developments in airborne LiDAR data acquisition have provided better horizontal and vertical ground information in the form of 3D point clouds. This has led to satisfactory results of LiDAR-derived digital terrain models (DTMs), also across complex ecosystems like natural forest stands. However, data and site-driven factors such as spatial resolution (point density), topography (slope and aspect), and variation in forest habitat types affect the DTM accuracy. In addition, processing steps like ground filtering and interpolation of ground points may also result in differences in DTM quality. Here, a comparative study was designed by extracting DTMs from two LiDAR data sources (high- and low-density point clouds) and three ground-filtering algorithms (adaptive TIN algorithm with and without the use of mirror points as well as an interpolation-based algorithm). The accuracy of the DTMs was assessed in association with terrain parameter and forest habitat types across heterogeneous forest sites of Bavarian Forest National Park in southeastern Germany. Qualitative analysis was carried out by taking 8300 independent sets of DGPS-recorded sample points. In addition to deriving root-mean-square error (RMSE) and bias, analysis of variance (ANOVA) type II was conducted in a factorial design to assess the influential factors on the observed DTM random error. Results revealed these errors in the DTMs with occasional over- and underestimations up to 1.98 m compared to reference elevation values. DTMs produced from high pulse density LiDAR data were more accurate than those extracted from low pulse density. Furthermore, topographic and forest habitat-type factors significantly contributed to the DTM accuracy. Slope increment showed a direct relationship with DTM error, with higher errors observed in south, southwest, and west aspects. Furthermore, stands dominated by deciduous trees were associated with higher DTM error than other forest habitat types. The applied adaptive TIN ground-filtering algorithms with mirror points and the interpolation-based algorithm both produced comparatively lower error rates, which are, therefore, suggested to reduce interpolation error in DTMs across rugged and heterogeneous forested terrains.

Keywords

LiDAR DTM Ground filtering Topography Forest habitat types Factorial analysis 

Zusammenfassung

Einfluss von Neigung, Exposition und Habitattyp auf die von LiDAR erfassten digitalen Geländemodelle bei naturnahen Mischwäldern in den gemäßigten Breiten. Airborne Laser Scanning (ALS, LiDAR) ermöglicht die Ableitung von hochpräzisen, horizontalen und vertikalen Geländeinformationen aus 3D-Punktwolken. Die daraus abgeleiteten digitalen Geländemodelle (DGM) sind von hoher Qualität, sogar in den komplexen Ökosystemen wie den heterogenen Waldbeständen Mitteleuropas. Allerdings wird die Genauigkeit eines DGMs zum einen durch Datengrundlage und Methoden, zum anderen durch standortbedingte Faktoren beeinflusst. Diese sind beispielsweise die räumliche Auflösung von LiDAR-Daten (Punktdichte), die Topographie (Hangneigung, Exposition) oder die Variabilität von Forsthabitaten. Darüber hinaus können Datenverarbeitungsschritte wie Filterung und Interpolation der Bodenpunkte zu Unterschieden in der DGM-Qualität führen. In dieser Studie wurden sechs DGMs verglichen, die aus zwei LiDAR-Datenquellen (hohe und niedrige Punktdichte) bestehen und mit drei Bodenfilteralgorithmen (Interpolation und adaptives TIN mit und ohne Spiegelpunkte) bearbeitet wurden. Die Genauigkeiten der DGMs wurden in Abhängigkeit von Topographiemerkmalen und Forsthabitattypen am Beispiel heterogener Waldgebiete des Nationalparks Bayerischer Wald untersucht. Die qualitative Analyse wurde mit Hilfe von 8300 voneinander unabhängigen Differenziellen GPS-Messungen durchgeführt. Die Einflussfaktoren auf den beobachteten Zufallsfehler wurden mit dem Root Mean Square Error (RMSE), der mittleren Abweichung (Bias). und zusätzlich mit einer Varianzanalyse ANOVA Typ II in einer Faktorenanalyse quantifiziert. Im Ergebnis wurden gelegentliche Über- und Unterschätzungen der Referenzhöhenwerte von maximal 1,98 m festgestellt. Die DGMs, welche aus hochpulsdichten LiDAR-Daten abgeleitet wurden, waren genauer als diejenigen, die aus Aufnahmen mit niedrigerer Pulsdichte stammten. Dabei wurde eine direkte Beziehung zwischen der Hangneigung und dem DGM-Fehler festgestellt. Die höchsten Fehlerraten wurden auf Hängen mit Süd-, Südwest-und West-Exposition beobachtet. Darüber hinaus waren im Vergleich zu anderen Lebensraumtypen die Laubbaumbestände mit höheren DGM-Fehlern assoziiert. Der adaptive TIN-Bodenfilter mit Spiegelpunkten sowie der interpolationsbasierte Algorithmus erzeugten vergleichsweise niedrigere Fehlerraten. Somit sind diese beiden Verfahren empfehlenswert, um LiDAR-basierte DGMs über heterogenen Waldökosystemen noch genauer abzuleiten.

Schlüsselwörter

LiDAR DGM Bodenfilterung Topographie Forsthabitat Faktorenanalyse 

Notes

Acknowledgements

The authors would like to thank the BFNP administration for providing LiDAR and field data used in this study. The RapidLasso GmbH provided a 3-month full licence for LiDAR data processing. The Bayerische Vermessungsverwaltung provided us with comprehensive information on flight and processing parameters for their DTM. The authors thank Prof. Florian Hartig for his support in terms of ANOVA type II analysis. In addition, we acknowledge the support of the Data Pool Initiative for the Bohemian Forest Ecosystem.

References

  1. Axelsson P (1999) Processing of laser scanner data-algorithms and applications. ISPRS J Photogramm Remote Sens 54:138–147. doi:10.1016/S0924-2716(99)00008-8 CrossRefGoogle Scholar
  2. Axelsson P (2000) DEM generation from laser scanner data using adaptive TIN models. Int Arc Photogramm Remote Sens 33(Part B4):110–117Google Scholar
  3. Bässler C, Hothorn T, Brandl R, Müller J (2013) Insects overshoot the expected upslope shift caused by climate warming. PLoS One 8(6):65842. doi:10.1371/journal.pone.0065842 CrossRefGoogle Scholar
  4. Chhatkuli S, Mano K, Kogure T, Tachibana K, Shimamura H (2012) Full waveform lidar exploitation technique and its evaluation in the mixed forest hilly region. ISPRS Int Arch Photogramm Remote Sens Spat Inf Sci XXXIX–B7(September):505–509Google Scholar
  5. Deems JS, Painter TH, Finnegan DC (2013) Lidar measurement of snow depth: a review. J Glaciol 59(215):467–479. doi:10.3189/2013JoG12J154 CrossRefGoogle Scholar
  6. Estornell J, Ruiz LA, Velázquez-Martí B, Hermosilla T (2011) Analysis of the factors affecting LiDAR DTM accuracy in a steep shrub area. Int J Digit Earth 4(6):521–538CrossRefGoogle Scholar
  7. Ewald M, Dupke C, Heurich M, Müller J, Reineking B (2014) LiDAR remote sensing of forest structure and GPS telemetry data provide insights on winter habitat selection of European roe deer. Forests 5(6):1374–1390CrossRefGoogle Scholar
  8. Fischer F, Knörzer O (2003) Statistische analyse von digitalen Geländemodellen und Waldstrukturen im Nationalpark Bayerischer Wald mit Hilfe von hochaufgelösten Laserscanningdaten und GPS-Messungen. Diploma thesis, University of Applied Sciences MunichGoogle Scholar
  9. Haneberg WC (2008) Elevation errors in a LIDAR digital elevation model of West Seattle and their effects on slope-stability calculations. Rev Eng Geol 20:55–65Google Scholar
  10. Hansen EH, Gobakken T, Næsset E (2015) Effects of pulse density on digital terrain models and canopy metrics using airborne laser scanning in a tropical rainforest. Remote Sens 7(7):8453–8468CrossRefGoogle Scholar
  11. Heurich M, Weinacker H (2004). Automated tree detection and measurements in temperate forest of central Europe using laser scanning data. ISPRS 2004 WG VIII/2, Oct. 3th–6th, University of FreiburgGoogle Scholar
  12. Heurich M, Fischer F, Knoerzeer O, Krzystek P (2008) Assessment of digital terrain models (DTM) from data gathered with airborne laser scanning in temperate European beech (Fagus sylvatica) and Norway spruce (Picea abies) forests. Photogramm Fenerkund Geoinf 6(2008):473–488Google Scholar
  13. Hodgson ME, Bresnahan P (2004) Accuracy of airborne LIDAR-derived elevation: Empirical assessment and error budget. Photogramm Eng Remote Sens 70(3):331–339CrossRefGoogle Scholar
  14. Hodgson ME, Jensen JR, Schmidt L, Schill S, Davis B (2003) An evaluation of LIDAR- and IFSAR-derived digital elevation models in leaf-on conditions with USGS Level 1 and Level 2 DEMs. Remote Sens Environ 84(2):295–308CrossRefGoogle Scholar
  15. Hyyppä J, Hyyppä H, Litkey P, Yu X, Haggrén H, Rönnholm P, Pyysal U, Pitkänen J, Maltamo M (2000) Algorithms and methods of airborne laser scanning for forest measurements. Int Arch Photogramm Remote Sens Spat Inf Sci 36(8):82–89Google Scholar
  16. Hyyppä H, Yu X, Hyyppä J (2005) Factors affecting the quality of DTM generation in forested areas. In: ISPRS WG III/3, III/4, V/3 workshop “Laser scanning 2005”, Enschede, The Netherlands, September 12–14, 2005Google Scholar
  17. Isenburg M (2013) LAS file processing using LASTOOLS 1–12Google Scholar
  18. Kraus K, Pfeifer N (2001) Advanced DTM generation from LIDAR data. Int Arch Photogramm Remote Sens XXXIV:22–24Google Scholar
  19. Kilian J, Haala N, Englich M (1996) Capture and evaluation of airborne laser scanner data. Int Arch Photogramm Remote Sens, vol. XXXI, Part B3. Vienna–AustriaGoogle Scholar
  20. Kraus K, Mikhail EM (1972) Linear least squares interpolation. Photogram Eng 38(10):1016–1029Google Scholar
  21. Kraus K, Pfeifer N (1998) Determination of terrain models in wooded areas with airborne laser scanner data. ISPRS J Photogramm Remote Sens 53(4):193–203CrossRefGoogle Scholar
  22. Krojerová-Prokešová J, Barančeková M, Šustr P, Heurich M (2010) Feeding patterns of red deer Cervus elaphus along an altitudinal gradient in the Bohemian Forest: effect of habitat and season. Wildl Biol 16(2):173–184CrossRefGoogle Scholar
  23. Latifi H, Fassnacht FE, Mueller J, Tharani A, Dech S, Heurich M (2015) Forest inventories by LiDAR data: a comparison of single tree segmentation and metric-based methods for inventories of a heterogeneous temperate forest. Int J Appl Earth Observ Geoinf 42:162–174CrossRefGoogle Scholar
  24. Lin X, Zhang J (2014) Segmentation-based filtering of airborne LiDAR point clouds by progressive densification of terrain segments. Remote Sens 6(2):1294–1326CrossRefGoogle Scholar
  25. Liu X (2008) Airborne LiDAR for DEM generation: some critical issues. Prog Phys Geogr 32(1):31–49CrossRefGoogle Scholar
  26. Maguya AS, Junttila V, Kauranne T (2013) Adaptive algorithm for large scale DTM interpolation from LiDAR data for forestry applications in steep forested terrain. ISPRS J Photogramm Remote Sens 85:74–83CrossRefGoogle Scholar
  27. McGhaughey R (2016) FUSION/LDV software for LiDAR data analysis and visualization. Pacific Northwest Research Station, USDA Forest Service, p 211. Accessed 27 Nov 2016Google Scholar
  28. Meng X, Currit N, Zhao K (2010) Ground filtering algorithms for airborne LiDAR data: a review of critical issues. Remote Sens 2(3):833–860CrossRefGoogle Scholar
  29. Peng M, Shih T (2006) Error assessment in two lidar-derived TIN datasets. Photogramm Eng Remote Sens 72(August):933–947CrossRefGoogle Scholar
  30. Pulighe G, Fava F (2013) DEM extraction from archive aerial photos: accuracy assessment in areas of complex topography. Eur J Remote Sens 46(1):363–378CrossRefGoogle Scholar
  31. Rapidlasso GMBH (2014) LAStools, “Efficient LiDAR Processing Software” (version 141017, academic) http://rapidlasso.com/LAStools. Accessed 27 Nov 2016
  32. Salleh MRM, Ismail Z, Rahman MZA (2015) Accuracy assessment of Lidar-derived digital terrain model (DTM) With different slope and canopy cover in tropical forest region. ISPRS Ann Photogramm Remote Sens Spat Inf Sci II–2/W2(October):183–189Google Scholar
  33. Selige T, Böhner J, Ringeler A (2006) Processing of SRTM X-SAR data to correct interferometric elevation models for land surface process applications. Göttinger Geographische Abhandlungen 115:97–104Google Scholar
  34. Silva CA, Hudak A (2013) Evaluation of digital elevation models (DEMs) from high and low pulse density in LiDAR data. In: Anias XVI symposium of remote sensing, Brazil, 13–18 April, pp 6065–6072Google Scholar
  35. Toz G, Erdogan M (2008) DEM (Digital Elevation Model) production and accuracy modeling of DEMS from 1:35.000 scale aerial photographs. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information 420 Sciences, vol. XXXVII, Part B1, Beijing, China, 6Google Scholar
  36. Yu X, Hyyppä H, Kaartinen H, Hyyppä J, Ahokas E, Kaasalainen S (2005) Applicability of first pulse derived digital terrain models for boreal forest studies. In: ISPRS WG III/3, III/4, 3(2004), pp 12–14Google Scholar
  37. Zhang S, Yang H, Singh L (2014) Increased information leakage from text. CEUR workshop proceedings 1225(February):41–42Google Scholar

Copyright information

© Deutsche Gesellschaft für Photogrammetrie, Fernerkundung und Geoinformation (DGPF) e.V. 2017

Authors and Affiliations

  1. 1.Department of Forest Research and SurveyKathmanduNepal
  2. 2.Department of Remote Sensing in Cooperation with German Aerospace Center (DLR)University of WürzburgWürzburgGermany
  3. 3.Department of Conservation and ResearchBavarian Forest National ParkGrafenauGermany
  4. 4.Laboratory for Engineering Measurement TechniquesHochschule für Technik StuttgartStuttgartGermany

Personalised recommendations