High-frequency over-the-horizon radar ship detection through clutter cancellation: an alternative to high-resolution spectral estimation

High-resolution spectral estimation is the resolution of spectral components that cannot be distinguished in the Fourier transform. This paper presents what is, in effect, an unusual techniques to perform high-resolution spectral estimation. By canceling powerful and obscuring ocean clutter, weak ship targets are exposed in the short-time Fourier transform of a particular kind of radar data. This achieves the same result as high-resolution spectral estimation, but it also preserves the desirable properties of the Fourier transform. A specific clutter-canceling algorithm has been developed and successfully tested on data from the US Navy's Relocatable Over-the-Horizon Radar. A side-by-side comparison of this algorithm with Tufts-Kumerisan forward-backward linear prediction, a popular high-resolution spectral estimator, will be presented. Although both techniques performed well at detecting the ship and estimating its Doppler frequency, the clutter cancellation techniques seems to provide a better estimate of the ship power relative to the peak of the clutter. This property would be useful in target identification.