Robust CFAR Detector Based on Truncated Statistics for Polarimetric Synthetic Aperture Radar

Abstract

Constant false alarm rate (CFAR) algorithms using local training window are widely used for ship detection with synthetic aperture radar (SAR) imagery. However, when the density of targets is high, such as in busy shipping lines and crowded harbors, the background statistics may be contaminated by the presence of nearby targets in the training window. Recently a robust CFAR detector based on truncated statistics (TS) was proposed. However, the truncation of data in the format of polarimetric covariance matrices is much more complicated with respect to the truncation of intensity (single polarization) data. In this paper a polarimetric truncated statistic CFAR (PWF-TS-CFAR) is proposed to estimate the background parameters accurately in the contaminated sea clutter for Pol SAR imagery. The CFAR detector uses polarimetric whitening filter (PWF) to turn the multidimensional problem to a one dimensional case. It uses the truncation to exclude possible statistically interfering outliers and uses TS to model the remaining background samples. The algorithm does not require a prior knowledge of the interfering targets and it is performed iteratively and adaptively to derive better estimates of the polarimetric covariance matrix (although this is computationally expensive). The PWF-TS-CFAR detector provides accurate background clutter modeling, a stable false alarm property, and improves detection performance in high-target-density situations. RadarSat2 data are used to verify our derivations and the results are in line with the theory.

Explore the application of high-resolution nighttime light remote sensing images in nighttime marine ship detection: A case study of LJ1-01 data

“…Among the many methods of marine ship detection, CFAR is a relatively excellent and commonly used algorithm [34][35][36]. With the development of research, the CFAR algorithm is gradually combined with machine learning methods such as convolutional neural networks and support vector machines, which further optimize the detection results [37][38][39].…”

A New Form of the Polarimetric Notch Filter

“…A CFAR framework called fast block CFAR is used [11]. The region of interest (ROI) is divided into equally sized blocks and then the brightest pixels are removed before estimation.…”

“…The CFAR loss and the FOM are shown in table 1 when the actual PFA is set 1e-5, and the generalized gamma distribution (GГD) is used to model the PNF and the NPNF. The parameter estimation is based on log-cumulants [4] [11]. The ROC curves are also shown in Fig 8 . The result from the ROC curves is in agreement with that from the FOM.…”

PolSAR Ship Detection Based on Neighborhood Polarimetric Covariance Matrix

“…It has been proven that polarimetry can provide more information in detecting and classifying targets in SAR images, leading to the wide application of polarimetric SAR (PolSAR) to improve the performances for small ships. A review of detection techniques is presented in [2], and recent advancements are shown in [3].…”