Improved Detection of the First-Order Region for Direction-Finding HF Radars Using Image Processing Techniques Academic Article uri icon

abstract

  • For direction-finding high-frequency (HF) radar systems, the correct separation of backscattered spectral energy due to Bragg resonant waves from that due to more complex double-scattering represents a critical first step toward attaining accurate estimates of surface currents from the range-dependent radar backscatter. Existing methods to identify this “first order” region of the spectra, generally sufficient for lower-frequency radars and low-velocity or low-surface gravity wave conditions, are more likely to fail in higher-frequency systems or locations with more variable current, wave, or noise regimes, leading to elevated velocity errors. An alternative methodology is presented that uses a single and globally relevant smoothing length scale, careful pretreatment of the spectra, and marker-controlled watershed segmentation, an image processing technique, to separate areas of spectral energy due to surface currents from areas of spectral energy due to more complex scattering by the wave field or background noise present. Applied to a number of HF radar datasets with a range of operating frequencies and characteristic issues, the new methodology attains a higher percentage of successful first-order identification, particularly during complex current and wave conditions. As operational radar systems continue to expand to more systematically cover areas of high marine traffic, close approaches to ports and harbors, or offshore energy installations, use of this type of updated methodology will become increasingly important to attain accurate current estimates that serve both research and operational interests.

publication date

  • August 2017