River embankment identification in the airborne laser scanning point cloud

• Autorzy: Tymków P. , Borkowski A. , Gołuch P.
• Języki publikacji: EN

Abstrakty

EN

Hydrodynamic modeling is one of the most significant tools in risk and environmental management of floodplains. Such research requires precise and reliable digital terrain models (DTM). Moreover, DTMs should determine the location of terrain edges for the area of river valley, i.e. river embankments. Terrain models are usually built using data collected with airborne laser scanning. Irregular cloud of scanned points can be used for 3D line edges modeling. However, this task is difficult because determination of the river embankment edges is not always exact; the edges are not defined precisely or they can be hidden in bushes and groups of higher vegetation. One of the newest methods of the edges modeling relies on the intersection of two planes. The main issue here is to classify points into subsets located on both sides of the modeled edge. In this study, the algorithm employing multilayer feed-forward neural network for point classification is presented. It allows us to include a priori information about the expected shape of surface as well as the orientation of embankment with respect to the river flow direction. Classification was performed on the real airborne laser scanning dataset. Confusion matrix was used for the quantitative accuracy assessment. This matrix was built for the test vector based on the comparison of the obtained results with an interactive sample.

Słowa kluczowe

PL

wał rzeczny; modelowanie hydrodynamiczne; skaning laserowy; cyfrowe modele terenu; DMT; modelowanie linii krawędziowych powierzchni

• Wydawca: De Gruyter
• Czasopismo: Studia Geotechnica et Mechanica
• Rocznik: 2009
• Tom: Vol. 31, nr 4
• Strony: 85--94
• Opis fizyczny: bibliogr. 10 poz.

Twórcy

autor

Tymków P.

autor

Borkowski A.

autor

Gołuch P.

• Institute of Geodesy and Geoinformatics, Wrocław University of Environmental and Life Sciences, ul. Grunwaldzka 53, 50-357 Wrocław,

Bibliografia

• [1] BRÜGELMANN R., Automatic breaklines detection from airborne laser scanner data, Int. Arch. Photogrammetry and Remote Sensing XXXIII(B3), 2000, 109–116.
• [2] SUI L., Ableitungtopographisher Strukturlinien aus Laserscannerdaten mit Methoden der Bildverarbeitung, Zeitschrift für Photogrammetrie, Fernerkundung, Geo-information, 2002, 423–434.
• [3] BRIESE C., Three-dimensional modelling of breaklines from airborne laser scanner data, ISPRS Congress, 12–23 July, Istanbul, Turkey, 2004, http://www.isprs.org/commission3/wg3
• [4] BORKOWSKI A., KELLER W., Global and local methods for tracking the intersection curve between two surfaces, Journal of Geodesy, Vol. 79, 2005, 1–10.
• [5] BORKOWSKI A., Modelowanie linii krawędziowych powierzchni na podstawie danych skaningu laserowego, Archiwum Fotogrametrii, Kartografii i Teledetekcji, Vol. 17, 2007, 73–82.
• [6] KASS M., WITKIN, A. TERZOPOULOS D., Snakes: Active contour models, Proceedings of the First International Conference of Computer Vision, IEEE Comput. Soc. Press, 1987, 259–268.
• [7] BORKOWSKI A., GOŁUCH P., WEHR A., SCHIELE O., THOMAS M., Airborne laser scanning for the purpose of hydrodynamic modelling of Widawa river valley, Reports on Geodesy, No 2(77), 2006, 85–94.
• [8] WEHR A., LOHR U., Airborne laser scanning – an introduction and overview, ISPRS Journal of Photogrammetry & Remote Sensing, 54, 1999, 68–82.
• [9] ADAMCZYK J., BĘDKOWSKI K., Metody cyfrowe w teledetekcji, Wydawnictwo SGGW, 2005.
• [10] KUBIK T., PALUSZYŃSKI W., IWANIAK A., TYMKÓW P., Klasyfikacja obrazów rastrowych z wykorzystaniem sztucznych sieci neuronowych i statystycznych metod klasyfikacji, Monografia, Wydawnictwo Uniwersytetu Przyrodniczego we Wrocławiu 2008.