A new robust method for bridge detection from high resolution electro optic satellite images

2012-05-09
Gedik, Ekın
Cınar, Umut
Karaman, Ersin
Çetin, Yasemin
Halıcı, Uğur
In this paper, an automatic approach for identifying bridges over water in satellite images is proposed. The proposed algorithm has three main steps. It starts with extracting the water regions in the satellite image by thresholding the NIR and clustering the NDWI images. Next possible river and water canals in the extracted water mask are identified by certain geometric constraints. Finally possible bridge regions are extracted by morphological operations applied to the water and river-canal mask. An optional step for verifying the bridge regions are also proposed in the study. The verification step uses orientation information or road mask for obtaining higher precision values. Tests are conducted using the proposed approach on different satellite images and quantitative and visual results showed that algorithm is successful in identifying different sized bridges over water.
Citation Formats
E. Gedik, U. Cınar, E. Karaman, Y. Çetin, and U. Halıcı, “A new robust method for bridge detection from high resolution electro optic satellite images,” Rio-De-Janeiro, Brezilya, 2012, p. 298, Accessed: 00, 2021. [Online]. Available: https://hdl.handle.net/11511/83279.