Fractal-multiresolution based detection of targets within sea clutter

Abstract

A wavelet transform focuses on localised signal structures with a zooming procedure that progressively reduces the scale parameter. On the other hand, fractal geometry has recently been applied to the analysis of high range resolution radar sea clutters. Using both concepts in designing a new detector, reveals considerable improvement in performance of target detection within sea clutter. In support of this argument, simulation results using real radar data samples are presented.Introduction: The concept of fractal geometry has been employed to model the roughness of sea surface and to investigate the scattering from rough surfaces [1]. Haykin et al.[2] applied fractal geometry theory to the analysis of real radar signals scattered from rough sea surfaces collected using IPIX radar. It was shown that the presence of a target changes the fractal dimension of the radar's returns and this variation could be used to detect the presence of a target. However, as stated in [2], the detection results are preliminary and are by no means conclusive at that stage of research. Salmasi and Modarres-Hashemi [3] demonstrated that the fractal detector has great capabilities in the rejection of coloured clutter. Nayebi and PourNejation [4] have not used the fractal dimension directly. Instead, correlation coefficients between the logarithm of box-counts and the scales in the time domain have been utilised. In our previous paper [5], an enhanced fractal based detection method was presented.In this Letter we propose a novel fractal based detector with multiresolution analysis, which takes into account the fractal dimension of the received signal. The role of this input is to provide finer analysis of the possible target echo embedded in a sea surface backscattered signal. Performance comparison of this detector and a fractal based [3] as well as energy detectors for real high resolution samples of sea clutter and noise shows significant improvement. The real data samples have been captured by IPIX radar [6].

Detection of small target in sea clutter via multiscale directional Lyapunov exponents

“…When the time scale is longer, such kind of method cannot detect the small target. To deal with the nonstationary sea clutter, the state-of-the-art methods are proposed, which include short-time fractional Fourier transform analysis technique (Chen et al , 2014), bi-window nonlinear shrinkage map-based approach (Xu et al , 2016), neural network-based approaches (Bhattacharya and Haykin, 1992; Bhattacharya and Haykin, 1997) (Leung et al , 2002), wavelet transform-based approach (Pournejatian and Nayebi, 2012) and wavelet-neural net combined approaches (Lin et al , 2000a; Lin et al , 2000b), multifractal analysis-based methods (Fan et al , 2015; Shi et al , 2016), etc. Besides the K distribution based methods, the current investigations suggest that in view of methodology, the multiscale analysis-based method should be adopted to reveal the local nonlinear dynamics of nonstationary time series and find a few computable features for accurately and easily detect targets within sea clutter.…”

Target Detection within Sea Clutter Using the Saturation Phenomenon of Fluctuation Analysis

“…It is known that the fractal theory provides an alternative way to describe sea clutter properties [1] and may effectively improve the algorithm performance for target detection [2]. For years, much effort has been committed to the analysis of sea clutter by fractal geometry [3]- [6]. Self-affinity is a kind of typical fractal characteristic.…”

A New Feature Based on the Fluctuation Analysis to Detect Low Observable Target within Sea Clutter

“…N.M. Pournejatian and M.M. Nayebi proposed a fractal-multiresolution detector for detecting low-observable targets within sea clutter [1]. Self-affinity is a typical generalized characteristic of fractals.…”