Automated correction of surface obstruction errors in digital surface models using off‐the‐shelf image processing

Airborne topographic data collection requires removal of errors that arise due to surface features that obstruct the ground from the sensor. Typically, this has been based on manual correction and/or automated filtering. To some degree, the latter has provided a method for identifying and removing unwanted surface obstructions in large topographic data-sets. However, the algorithms used are unintelligent in that they cannot reliably differentiate between the various types of obstructions and the ground. If coincident optical support imagery is available, the use of intelligent correction routines becomes possible. This paper describes an automated approach for removing obstruction errors using optical support imagery and simple image processing routines. Orthorectification and classification of support imagery enable obstruction errors to be identified in the digital surface model (DSM) and corrected intelligently to produce a digital terrain model (DTM). The results show that support imagery can be used with basic image processing routines to remove obstructions intelligently and automatically from large topographic data-sets. Since the approach can differentiate between types of obstructions, the removal of each type of error can be customised, making this a very flexible approach to topographic data correction. Resume a saisie aeroportee des donnees topographiques doit etre suivie de l’elimination des erreurs dues au masquage du terrain par des objets du sur-sol. Cela s'obtient soit par des corrections manuelles soit par un filtrage automatique ou par les deux. Le filtrage automatique a permis, jusqu’a un certain degre, d'identifier et d’eliminer ces obstructions de surface indesirables pour la construction des grands jeux de donnees topographiques. Toutefois les algorithmes utilises n'ont aucune intelligence en ce sens qu'ils ne permettent pas d'identifier valablement les differents types d'obstruction et de les differencier du sol. Si l'on dispose aussi d'une imagerie sur support optique, il devient alors possible d'utiliser des programmes intelligents de correction. On presente dans cet article une solution automatique d’elimination des erreurs de masquage basee sur l'emploi d'images sur support optique et de logiciels simples de traitement d'images. Le redressement differentiel et la classification d'images permettent d'identifier les erreurs de masquage dans le modele numerique de surface (MNS) et de les eliminer intelligemment pour aboutir au modele numerique de terrain (MNT). Les resultats ont montre que l'on pouvait effectivement utiliser cette imagerie avec des logiciels de traitement d'images basiques pour eliminer automatiquement et intelligemment ces obstructions dans les grandes bases de donnees topographiques. Etant donne que cette solution permet de differencier les divers types d'obstructions, on peut personnaliser a la demande l’elimination de chaque type d'erreur et obtenir ainsi un moyen tres souple de correction des donnees topographiques. Zusammenfassung Die Erfassung der Topographie mittels digitaler Oberflachenmodelle erfordert die Eliminierung von Fehlern, die durch Oberflachenmerkmale hervorgerufen werden, die die Gelandeoberflache verdecken. Typischerweise wird dies durch manuelle Editierung und/oder automatische Filterverfahren erreicht. Bis zu einem gewissen Grad sind letztere geeignet, in grosen topographischen Datensatzen unerwunschte 3D-Strukturen zu identifizieren und zu beseitigen. Dennoch sind die Algorithmen nicht in der Lage, zuverlassig zwischen verschiedenen Arten von Verdeckungen und der Gelandeoberflache zu unterscheiden. Sind jedoch korrespondierende optische Bilddaten verfugbar, konnen geeignetere Korrekturverfahren eingesetzt werden. In diesem Beitrag wird ein automatisierter Ansatz zur Beseitigung von 3D-Strukturen in digitalen Oberflachenmodellen beschrieben, in dem zusatzlich einfache Bildverarbeitungsroutinen auf korrespondierende optische Bilddaten angewandt werden. Eine Orthobildentzerrung und Klassifizierung der Bilddaten erlaubt die Erkennung und Eliminierung der Strukturfehler im digitalen Oberflachenmodell (DSM) und die Ableitung eines digitalen Gelandemodells (DTM). Die Ergebnisse bestatigen, dass dieser Ansatz zur automatischen Beseitigung von 3D-Strukturfehlern in grosen topographischen Datensatzen geeignet ist. Da hier zwischen Typen von Strukturfehlern unterschieden werden kann, ist es moglich, die Bereinigung individuell anzupassen, was eine sehr flexible Korrektur topographischer Daten erlaubt. Resumen La captura aerea de informacion topografica requiere eliminar los errores que surgen a causa de los elementos de la superficie que apantallan el terreno respecto del sensor. Comunmente se ha procedido mediante correccion manual y/o filtrado automatico. En cierto modo, este ultimo metodo ha permitido identificar y eliminar los errores indeseados producidos por obstrucciones en la superficie contenidos en conjuntos de datos topograficos muy grandes. Sin embargo, los algoritmos usados no son inteligentes, ya que no pueden distinguir con fiabilidad entre los diferentes tipos de obstaculos y el terreno. Si se dispone de imagenes opticas de apoyo, es posible utilizar rutinas de correccion inteligente. Este articulo describe un procedimiento automatizado para eliminar los errores por obstaculos usando imagenes opticas de apoyo y simples rutinas de procesamiento de imagenes. La ortorectificacion y la clasificacion de imagenes de apoyo permiten identificar los errores producidos por obstaculos en el modelo digital de la superficie y corregirlos inteligentemente para calcular un modelo digital del terreno. Los resultados indican que las imagenes de apoyo pueden usarse conjuntamente con rutinas basicas de proceso de imagenes para eliminar las obstrucciones de forma automatica e inteligente. Dado que el procedimiento sabe diferenciar entre tipos de obstruccion, puede personalizarse la eliminacion de cada tipo de error, lo que hace que este metodo sea muy flexible para la correccion de datos topograficos.

[1]  L. M Gomes Pereira,et al.  Suitability of laser data for deriving geographical information: A case study in the context of management of fluvial zones , 1999 .

[2]  T. M. Lillesand,et al.  Remote Sensing and Image Interpretation , 1980 .

[3]  Stephen Wise,et al.  Assessing the quality for hydrological applications of digital elevation models derived from contours , 2000 .

[4]  P. Mather,et al.  Classification Methods for Remotely Sensed Data , 2001 .

[5]  Richard C. Dubes,et al.  Performance evaluation for four classes of textural features , 1992, Pattern Recognit..

[6]  M.J.P.M. Lemmens Accurate height information from airborne laser-altimetry , 1997, IGARSS'97. 1997 IEEE International Geoscience and Remote Sensing Symposium Proceedings. Remote Sensing - A Scientific Vision for Sustainable Development.

[7]  P. Swain,et al.  Neural Network Approaches Versus Statistical Methods In Classification Of Multisource Remote Sensing Data , 1990 .

[8]  Gabriele Bitelli,et al.  Comparison of Techniques for Generating Digital Terrain Models from Contour Lines , 1997, Int. J. Geogr. Inf. Sci..

[9]  P. Reiss,et al.  Laser scanning—surveying and mapping agencies are using a new technique for the derivation of digital terrain models , 1999 .

[10]  John R. Jensen,et al.  Introductory Digital Image Processing: A Remote Sensing Perspective , 1986 .

[11]  S. N. Lane,et al.  Application of Digital Photogrammetry to Complex Topography for Geomorphological Research , 2000 .

[12]  S. Lane,et al.  The Measurement of River Channel Morphology Using Digital Photogrammetry , 2000 .

[13]  Stuart N. Lane,et al.  REMOTE SENSING OF CLEAR-WATER, SHALLOW, GRAVEL-BED RIVERS USING DIGITAL PHOTOGRAMMETRY , 2001 .

[14]  George Vosselman,et al.  Two algorithms for extracting building models from raw laser altimetry data , 1999 .

[15]  Stuart N. Lane,et al.  The development of an automated correction ­procedure for digital photogrammetry for the study of wide, shallow, gravel‐bed rivers , 2000 .

[16]  John Trinder,et al.  Using the Dempster-Shafer method for the fusion of LIDAR data and multi-spectral images for building detection , 2005, Inf. Fusion.

[17]  David C. Mason,et al.  Obtaining accurate maps of topography and vegetation to improve 2D hydraulic flood models , 2001, SPIE Remote Sensing.

[18]  P. Bates,et al.  Integration of high-resolution topographic data with floodplain flow models. , 2000 .

[19]  Claus Brenner,et al.  Extraction of buildings and trees in urban environments , 1999 .

[20]  Martin J. Wooster,et al.  Texture based feature extraction: Application to burn scar detection in Earth observation satellite sensor imagery , 2002 .

[21]  F. Ackermann Airborne laser scanning : present status and future expectations , 1999 .

[22]  Rajendra P. Shrestha,et al.  Application of DEM Data to Landsat Image Classification: Evaluation in a Tropical Wet-Dry Landscape of Thailand , 2000 .

[23]  Michael Michael Hord,et al.  Digital image processing of remotely sensed data , 1982 .

[24]  John R. G. Townshend,et al.  Terrain Analysis and Remote Sensing , 1981 .

[25]  Aloysius Wehr,et al.  Airborne laser scanning—an introduction and overview , 1999 .

[26]  S. Lane,et al.  Remote survey of large-scale braided, gravel-bed rivers using digital photogrammetry and image analysis , 2003 .

[27]  Lotfi A. Zadeh,et al.  Fuzzy Sets , 1996, Inf. Control..

[28]  Emmanuel P. Baltsavias,et al.  A comparison between photogrammetry and laser scanning , 1999 .

[29]  Peter Axelsson,et al.  Processing of laser scanner data-algorithms and applications , 1999 .

[30]  S. Lane,et al.  Estimation of erosion and deposition volumes in a large, gravel‐bed, braided river using synoptic remote sensing , 2003 .

[31]  D. Mason,et al.  Image processing of airborne scanning laser altimetry data for improved river flood modelling , 2001 .