Home // International Journal On Advances in Intelligent Systems, volume 6, numbers 1 and 2, 2013 // View article

Horizon Line Detection in Marine Images: Which Method to Choose?

Authors:
Evgeny Gershikov
Tzvika Libe
Samuel Kosolapov

Keywords: horizon detection; marine images; edge detection; median filtering; image analysis; local edge magnitudes; regional edge magnitudes; regional covariances

Abstract:
Five algorithms designed to find a horizontal line separating sea and sky in marine images in real-life conditions were implemented in this work and compared by their performance: accuracy and relative speed. One of the selected algorithms was based on regional covariances in luminance images, the second one was based on edge detection and the Hough transform, the third one used maximal local edge detection and the least-squares method, the fourth one employed median filtering in small neighborhoods and linear regression and the fifth one was based on regional edge magnitudes and the least-squares method. Real-life images were used for comparison. The most accurate line with respect to the angular error was obtained by using the edge detection and Hough transform based algorithm. The highest accuracy with respect to the position of the line was achieved by the regional covariance method. However, the highest speed was achieved by using the regional edge magnitude algorithm.

Pages: 79 to 88

Copyright: Copyright (c) to authors, 2013. Used with permission.

Publication date: June 30, 2013

Published in: journal

ISSN: 1942-2679