Curvelet transform for water bodies extraction from high resolution satellite images

Document Type : Original Article

Authors

1 Ain Shams University, Cairo.

2 Egyptian Armed Forces.

3 Dept. of Geomatics Engineering, University of Calgary, Calgary, Alberta.

Abstract

In this research paper, a new implementation on the second generation curvelet transform in the edge detection of coastline is presented and applied on WorldView-2 imagery, together with a comparison with the classical edge detection methods such as Canny operator and the traditional wavelet transforms. This implementation is aiming to compare this new approach to the traditional edge detection techniques. It is found that the curvelet proposed implementation performs better in detecting larger and elongated structures compared to the Canny and the wavelet transforms. However, Although this method is promising and efficient for edge detection, the quality of the edge detection is still a function of the pre-processing steps (the classification step in this research paper) , as any edge detector will suffer from the heterogeneity of the images especially when using very high resolution imagery.

Keywords

Files

Share

How to cite

Statistics