Landmark detection from sidescan sonar images
Authors: Al-Rawi, M., Galdran, A., Elmgren, F., Rodriguez, J., Bastos, J. and Pinto, M.
Journal: 2017 IEEE Jordan Conference on Applied Electrical Engineering and Computing Technologies, AEECT 2017
Volume: 2018-January
Pages: 1-6
ISBN: 9781509059690
DOI: 10.1109/AEECT.2017.8257760
Abstract:Sidescan sonars have seen wide deployment in underwater imaging. They can be used to image the seabed to a rather acceptable resolution from a few centimeters to 10 centimeters. Yet, sonar images are still of a substantially lower visual quality as they suffer from quite a few problems, e.g., acoustic shadows that vary according to vehicle heading and sonar grazing angle, speckle noise, geometric deformation due to ping variation and speed of vehicle carrying the sonar, etc. Landmark detection in sidescan sonar images is vital to find objects and locations of interest that are useful in various underwater operations. The objective of this work is proposing novel landmark detection methods for this class of images. Cubic smoothing spline fitted to the across-Track signals is proposed as a method to detect the objects and their shadows. To cover a large area, experimental data has been acquired during missions performed in Melenara Bay (Las Palmas/Spain) using autonomous underwater vehicles (AUVs) equipped with Klein 3500 sidescan sonar. The AUVs have been deployed in two missions (one mission performed each day) and a total of 25 large-resolution images have been acquired. The AUV generated 12 parallel path images in the first mission and 13 parallel path images in the second mission with an angle of 70 degrees between the direction of mission #1 and mission #2. This difference in the directions of the two missions was necessary to ensure different acoustic shadows between the two sets of images, each set being generated from a different mission. Results show that the proposed methods are powerful in detecting landmarks from these challenging images.
Source: Scopus
Landmark detection from sidescan sonar images
Authors: Al-Rawi, M., Galdran, A., Elmgren, F., Rodriguez, J., Bastos, J. and Pinto, M.
Journal: 2017 IEEE JORDAN CONFERENCE ON APPLIED ELECTRICAL ENGINEERING AND COMPUTING TECHNOLOGIES (AEECT)
ISSN: 2378-4903
Source: Web of Science (Lite)