Video recognition of breaking waves

Author

Rusch, Christian ; Thomson, J. ; Zippel, S. ; Schwendeman, M.

Author_Institution

Appl. Phys. Lab., Univ. of Washington, Seattle, WA, USA

fYear

2014

fDate

14-19 Sept. 2014

Firstpage

1

Lastpage

6

Abstract

An algorithm is presented to automate the identification of breaking waves in images collected with a camera on a drifting buoy. Each image is given a score from four separate analysis techniques: brightness detection, pixel histogram, entropy (texture) analysis, and glare identification. By combining these in a composite score, potential breaking wave images are detected and the number of images requiring manual review is a small fraction of the original set. Most of the images with false breaking wave signals due to sun glare are identified and removed. The final output is the wave-breaking rate over the length of the video capture.

Keywords

entropy; ocean waves; oceanographic techniques; analysis technique; breaking wave identification automation; breaking wave video recognition; brightness detection; composite score; drifting buoy; entropy analysis; false breaking wave signal; glare identification; image manual review number; pixel histogram; potential breaking wave image detection; sun glare; video capture length; wave-breaking rate final output; Algorithm design and analysis; Cameras; Heuristic algorithms; Histograms; MATLAB; Manuals; Vectors; Wave-breaking; image processing;

fLanguage

English

Publisher

ieee

Conference_Titel

Oceans - St. John's, 2014

Conference_Location

St. John´s, NL

Print_ISBN

978-1-4799-4920-5

Type

conf

DOI

10.1109/OCEANS.2014.7002994

Filename

7002994