Coastline Extraction From High-Resolution Multispectral Images by Integrating Prior Edge Information With Active Contour Model

Abstract

Monitoring coastlines is essential for scientific protection and utilization of coastal zones. Since it is very common that various types of coasts are intermixed, there is much practical significance to propose a method that has universality. Moreover, weak edges and sea noise are the main problems of coastline delineation with high-resolution multispectral imagery. To settle these problems, a new method integrating prior edge information with an edge-based active contour model (EI-ACM) is proposed in this article. First, the coarse sea region is obtained with a novel region growth method. An edge refining process is then proposed to refine the edges detected by the Canny operator and obtain the prior edge information. The prior edge information is subsequently integrated with an EL-ACM to guide its evolution. Finally, the coastline is obtained by contour evolution. To validate the proposed method, it was compared to the state-of-the-art coastline extraction method. The results proved that the proposed EI-ACM outperforms in weak edge retention and avoiding the interference of sea noise, while being robust to various type of coastal.