Change detection of ocean wave characteristics. (1st June 2016)

• Record Type: Journal Article
• Authors: Zhu, Nuoyi; Kim, Yeesock; Kim, Kyu-Han; Shin, Bum-Shick
• Is Part Of: Expert systems with applications. Volume 51(2016)
• Page Start: 245
• Page End: 258
• Publication Date: 2016-06-01
• Subjects: Fuzzy C-means clustering -- Feature extraction -- Short-time Fourier transform (STFT) -- High frequency (HF) radar -- Ocean wave height -- Ocean wave spectrum; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2015.12.017
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: The first attempt to explore feature extractions of ocean HF radar data. Developed the systematic feature extraction algorithm for complex ocean wave signals. The high performances on the characteristic change detection of ocean waves. Abstract: In this paper, a novel feature extraction approach is proposed for identifying ocean wave characteristics in real time. The algorithm was developed through the integration of the fuzzy C-means clustering algorithm, statistics formulation, short-time Fourier transforms, high frequency radar data processing and window function analysis. This method provides new insight into the detection of ocean wave characteristics and provides a more direct and convenient way to detect changes in ocean wave characteristics than the conventional method. To demonstrate the proposed algorithm, two Wellen radar systems were installed in Samcheok City, Gangwon-do on the East Coast of South Korea. A data set was selected for training the proposed algorithm while three other data sets, not used for the training processes, were used to validate the proposed model. The testing results demonstrate that the proposed algorithm is effective in extracting characteristic features from a variety of ocean waves. It is expected that the proposed system will accurately predict natural hazards and provide adequate warning time for people to evacuate from threatened coastal area. Hence this approach will directly contribute to the reduction of injuries and deaths in natural disasters by supplying near real-time data of the environment around coastal areas.