Target Detection in Sea Clutter with Application to Spaceborne SAR Imaging

TL;DR

This work addresses target detection in sea clutter for spaceborne SAR by identifying an accurate statistical model of background clutter. Using KL divergence as a goodness-of-fit test, the authors show the Weibull distribution best describes sea clutter among several candidates and then develop a 2D CFAR detector whose adaptive threshold $T_{aW}=\u03b2[\log(1/p_{fa})]^{1/\u03b1}$ is tuned via ML estimates $\hat{\u03b1},\hat{\u03b2}$. The approach is validated on real RADARSAT-1 data from English Bay, Vancouver, with Range-Doppler reconstruction, Doppler centroid estimation, and 2D CFAR detection that successfully identifies ships amidst clutter, aided by speckle suppression. The combination of KL-based model selection and Weibull CFAR yields robust ship detection in challenging sea clutter, offering practical improvements for remote sensing and surveillance applications. Overall, the paper provides a principled, data-driven method for adaptive target detection in sea clutter using spaceborne SAR data.

Abstract

In this paper, the challenging task of target detection in sea clutter is addressed. We analyze the statistical properties of the signals which have been received from the scene and based on that, we model the amplitude of the signals that have been reflected from the background sea clutter according to several well-known probability distribution functions. Next, by exploiting the Kullback-Leibler (KL) divergence metric as a goodness-of-fit test, we will demonstrate that among the proposed probability distributions, the Weibull distribution can model the statistical properties of the background sea clutter with higher accuracy. Subsequently, we utilize the aforementioned information to design an adaptive threshold based on the Constant False Alarm Rate (CFAR) algorithm to detect the energy of the targets which have been buried in the sea clutter. Thorough analysis of the experimental data gathered from the Canadian RADARSAT-1 satellite demonstrates the overall effectiveness of the proposed method.

Figures (8)

• Figure 1: The range compressed data for several strong isolated targets. In this image, the isolated targets are several ships in English Bay.
• Figure 2: The amplitude-based estimation of the fractional part of the Doppler centroid frequency with its maximum value at $f^{\prime}_{\rm dc} = 531 \; \rm Hz$.
• Figure 3: The reconstructed image from English Bay, in Vancouver Canada, based on the Range-Doppler algorithm.
• Figure 4: The result of the speckle noise reduction process, based on (\ref{['speckle']}), for the reconstructed image of English Bay, in Vancouver Canada, shown in Fig. \ref{['fig:EnglishBay_Speckle']}.
• Figure 5: The histogram of the experimental data inside the red box which is related to the sea clutter as well as the estimated probability density functions based on (\ref{['Weibull']}), (\ref{['lognormal']}), (\ref{['InverseGaussian']}), (\ref{['Gamma']}), and (\ref{['Rayleigh']}). The $I$ and $Q$ represent the imaginary and quadratic part of the estimated value for the complex radar cross section.