Desmond Power

Target detection in synthetic aperture radar imagery: a state-of-the-art survey

Abstract Target detection is the front-end stage in any automatic target recognition system for synthetic aperture radar (SAR) imagery (SAR-ATR). The efficacy of the detector directly impacts the succeeding stages in the SAR-ATR processing chain. There are numerous methods reported in the literature for implementing the detector. We offer an umbrella under which the various research activities in the field are broadly probed and taxonomized. First, a taxonomy for the various detection methods is proposed. Second, the underlying assumptions for different implementation strategies are overviewed. Third, a tabular comparison between careful selections of representative examples is introduced. Finally, a novel discussion is presented, wherein the issues covered include suitability of SAR data models, understanding the multiplicative SAR data models, and two unique perspectives on constant false alarm rate (CFAR) detection: signal processing and pattern recognition. From a signal processing perspective, CFAR is shown to be a finite impulse response band-pass filter. From a statistical pattern recognition perspective, CFAR is shown to be a suboptimal one-class classifier: a Euclidean distance classifier and a quadratic discriminant with a missing term for one-parameter and two-parameter CFAR, respectively. We make a contribution toward enabling an objective design and implementation for target detection in SAR imagery.

Automatic Target Recognition in Synthetic Aperture Radar Imagery: A State-of-the-Art Review

The purpose of this paper is to survey and assess the state-of-the-art in automatic target recognition for synthetic aperture radar imagery (SAR-ATR). The aim is not to develop an exhaustive survey of the voluminous literature, but rather to capture in one place the various approaches for implementing the SAR-ATR system. This paper is meant to be as self-contained as possible, and it approaches the SAR-ATR problem from a holistic end-to-end perspective. A brief overview for the breadth of the SAR-ATR challenges is conducted. This is couched in terms of a single-channel SAR, and it is extendable to multi-channel SAR systems. Stages pertinent to the basic SAR-ATR system structure are defined, and the motivations of the requirements and constraints on the system constituents are addressed. For each stage in the SAR-ATR processing chain, a taxonomization methodology for surveying the numerous methods published in the open literature is proposed. Carefully selected works from the literature are presented under the taxa proposed. Novel comparisons, discussions, and comments are pinpointed throughout this paper. A two-fold benchmarking scheme for evaluating existing SAR-ATR systems and motivating new system designs is proposed. The scheme is applied to the works surveyed in this paper. Finally, a discussion is presented in which various interrelated issues, such as standard operating conditions, extended operating conditions, and target-model design, are addressed. This paper is a contribution toward fulfilling an objective of end-to-end SAR-ATR system design.

Ocean clutter suppression for an HF ground wave radar

At HF frequencies, the sea echo or clutter return is quite significant and can mask signals from targets travelling at or near the same Doppler of the resonant (Bragg) ocean waves. This paper presents a systematic method of reducing clutter signals so as to enable detection of targets travelling at speeds near the Bragg Doppler. The algorithm uses a Hankel rank reduction method that assumes that targets and ocean clutter can be modelled as time varying sinusoids. Several examples of clutter suppression are shown from data that was collected at the Cape Race facility. A cooperating vessel was used to show that a clutter supression algorithm may be used to enhance target detection when the targets speed is near the Bragg Doppler frequency.

Comparison of C-band SAR algorithms to derive surface wind vectors and initial findings in their use in marine search and rescue

This paper presents a comparison of existing algorithms to derive surface winds from synthetic aperture radar (SAR) satellites and investigates their use in drift forecasting for search and rescue purposes. Specifically, SAR-derived winds from RADARSAT-1 and ENVISAT advanced synthetic aperture radar (ASAR) data at 1.5 km resolution are compared with scatterometer-derived winds. Three approaches were used to retrieve the wind vector from the SAR data, including an optimal inversion method combining SAR data and background numerical weather prediction, the geophysical model function CMOD-IFR2 with an a priori wind direction, and a technique that uses the backscatter values corresponding to two neighboring subimages with slightly different incidence angles. Our comparisons of SAR wind mapping with scatterometer winds from QuikSCAT and ERS-2 produced a root mean square error (RMSE) of 1.5 m/s. The optimal inversion method seems very promising and appears to be the best choice for assimilation of SAR-derived winds into operational wind products with respect to the datasets presented here. Additionally, the suitability of SAR imagery for search and rescue operations is reviewed. It is recommended that a method should be explored to automatically assimilate such data into operational search and rescue tools. Use of SAR winds in a search and rescue drift model is shown herein to produce improved drift trajectories on a number of search and rescue targets (e.g., life boat, sail boat, person in water).

Iceberg and ship discrimination with ENVISAT multipolarization ASAR

Spaceborne synthetic aperture radar (SAR) can provide wide area and all-weather surveillance for iceberg and ship targets. However, the discrimination between icebergs and ships in SAR imagery, especially in the single polarization imagery that has been available over the past decade, is not always reliable. This is especially true when vessel and iceberg size are on the order of the pixel spacing. Present requirements for ocean surveillance with SAR data include a high detection and classification accuracy due to the necessity of comparable performance with other reconnaissance methods, such as aerial. ENVISAT advanced SAR (ASAR) data offers a potential solution to the iceberg-ship discrimination problem. ASAR data has comparable swath and resolution to other operational SAR systems and in addition offers an alternating polarization (AP) mode. AP targets offer more information than single polarization with respect to radar scattering mechanisms. The AP ship and iceberg targets in this study were observed to have considerably different polarization responses. In particular, ship targets in the HH and HV channels were comparable. In contrast, iceberg targets had at best, weak HV responses compared to the HH channel. Two methods for target discrimination were investigated: a multipolarized area ratio and HV signal-to-clutter ratio (SCR).

Characterization and statistical modeling of phase in single-channel synthetic aperture radar imagery

Traditionally, the phase content in single-channel synthetic aperture radar (SAR) imagery is discarded. This practice is justified by conventional radar resolution theory, which is a theory strictly relevant to point targets. The advent of high-resolution radars permits small targets previously considered to be points to be now treated as extended targets, in which case this theory is not strictly applicable. With this in mind, this paper offers a new insight into the relevance of phase in single-channel SAR imagery. The proposed approach builds on techniques from the fields of complex-valued and directional statistics. In doing so, three main contributions are presented, the first being a novel method for characterizing the phase content. Second, a new statistical model for the phase is considered, and then a set of 15 solely phase-based features are discussed. Our results are demonstrated on real-world SAR datasets for ground-truthed targets. The statistical significance of the information carried in the phase is clearly demonstrated. Furthermore, if applied to a dataset with higher resolution, the proposed techniques are expected to achieve even higher performance.

Well site extraction from Landsat-5 TM imagery using an object-and pixel-based image analysis method

Well sites, including both well pads and exploratory core holes, are small polygonal landscape disturbance features approximately one half to one hectare (0.5–1 ha) in area, resulting from oil and gas exploration activities. Automatic extraction and monitoring of such small features using remote-sensing technology at regional scales has always been desirable for wildlife habitat monitoring and environmental planning and modelling. Due to the vast disturbances of well sites in a province like Alberta, Canada, high-resolution imagery is not practical for well site extraction. For operational purposes, mid-resolution and cost-effective satellite imagery such as Landsat is the choice. However, automatic well site extraction using mid-resolution satellite imagery is a challenging task. Wells are typically less than three pixels in width and length in a Landsat multispectral image. Furthermore, the spectral contrast between the well site pixels and the surrounding areas is low due to vegetation regrowth and the spectral complexity of the surrounding environment. This article presents a novel methodology for automatic extraction of well sites from Landsat-5 TM imagery. The method combines both pixel- and object-based image analyses and contains three major steps: geometric enhancement, segmentation, and well site extraction. The method was applied to Landsat-5 TM images acquired over Fort McMurray, Alberta, Canada. For accuracy assessment, four regions of interest were selected and the results of the proposed automatic method were evaluated against visual inspection of the Landsat-8 pan-sharpened image. The method results in a total average correctness, completeness, and quality measures of about 80, 96, and 77%, respectively over the four sites. In addition, the method is very fast as an entire Landsat scene is processed in less than 10 minutes. The method is an operational approach for automatic detection of well sites over the entire province and can dramatically reduce the labour cost of manual digitization for monitoring and updating well site maps.

Dual Polarization Detection of Ships and Icebergs - Recent Results with ENVISAT ASAR and Data Simulations of RADARSAT-2

The RADARSAT-2 satellite is an advanced C-band synthetic aperture radar (SAR) with a variety of new modes including options for polarization combinations, resolution, and swath width. This paper examines the potential of multi polarization data for detecting and discriminating ship and iceberg targets Data used in this study consist of well validated airborne Convair-580 SAR and spaceborne ASAR HH/HV and HH/VV. In total, the data set used for evaluating detection and discrimination consists of 901 validated iceberg and ship targets. Optimizing target detection is accomplished using receiver operator curves (ROC) as proposed by [6] and discrimination is conducted using a quadratic discriminant (QD) with feature selection based on sequential forward selection (SFS). In general it was found that detection and discrimination improve with more polarimetric information; however, HH/HV and VV/VH only had nominally less discrimination performance than the quad polarization modes evaluated.

Nonlinearity and the effect of detection on single-channel synthetic aperture radar imagery

When signals exhibit non-Gaussian statistics, nonlinear signal processing techniques offer advantages over their linear counterparts. Nonlinearity in high-resolution synthetic aperture radar (SAR) imagery is an intrinsic phenomenon often overlooked in the radar literature. In this paper, we study the nonlinear dynamics, and the effect of detection, in SAR imagery. To this end, two complementary methods for exposing the nonlinear statistics are presented. The first method utilizes histogram fitting with relevant statistical models. The second method is based on hypothesis testing. Our results are demonstrated on real-world Radarsat-2 target chips. It is found that in the presence of extended targets (e.g., ships), the nonlinear effect in the SAR chip is predominant. Nonlinearity is observed to be negligible in the absence of extended targets. As the SAR chip is detected, the nonlinear dynamics are either diminished/altered (i.e., for power-detection) or obliterated (i.e., for magnitude-detection). To take full advantage of nonlinear statistics, it is recommended to utilize the complex-valued SAR image rather than the detected one. Furthermore, the Students T location-scale distribution is seen to offer an excellent model for the SAR chip.

On circularity/noncircularity in single-channel synthetic aperture radar imagery

Motivated by the conventional resolution theory, phase content in single-channel synthetic aperture radar (SAR) imagery is often discarded. In this paper, the validity of this practice is studied from the perspective of complex-valued statistics. Hence, for the phase content to be irrelevant, the complex-valued random variable has to be second-order circular. A procedure to characterize circularity/noncircularity in single-channel SAR imagery is presented. Our analysis is applied to real-world SAR chips from Radarsat-2 and MSTAR. For the case of extended targets, the complex-valued SAR chip is found to be inherently noncircular. Further, the strength of noncircularity is observed to be resolution-dependent. Also, a proportional relationship between noncircularity and nonlinearity is noted. These findings warrant investigating the statistical significance of this phenomenon in relevant target recognition applications.