Publications

Various technologies for mitigating interference in stretch-processed SAR imagery are described herein. Stretch-processed SAR data is received at a computing device. The stretch-processed (or deramped) SAR data is then reramped, thereby removing frequency-variant components of narrowband interference signals in the deramped data. A frequency-domain transform is executed over the reramped data to generate a spectral characteristic of the reramped data. A spectral notch filter is applied to frequency bands corresponding to the peaks of the spectral characteristic in order to filter out the narrowband interference signals. An inverse frequency-domain transform can then be executed over the filtered spectral characteristic to return to a phase-history representation of the SAR data. The phase history resulting from the inverse frequency-domain transform is a ramped phase history, which can then be deramped prior to use in connection with generating images of the scanned scene.

For synthetic aperture radar systems, missing data samples can cause severe image distortion. When multiple, coherent data collections exist and the missing data samples do not overlap between collections, there exists the possibility of replacing data samples between collections. For airborne radar, the known and unknown motions of the aircraft prevent direct data sample replacement to repair image features. This paper presents a method to calculate the necessary phase corrections to enable data sample replacement using only the collected radar data.

Missing samples within synthetic aperture radar data result in image distortions. For coherent data products, such as coherent change detection and interferometric processing, the image distortion can be devastating to these second-order products, resulting in missed detections, and inaccurate height maps. Previous approaches to repair the coherent data products focus upon reconstructing the missing data samples. This paper demonstrates that reconstruction is not necessary to restore the quality of the coherent data products.

Abstract not provided.

Synthetic aperture radar (SAR) images contain a grainy pattern, called speckle, that is a consequence of a coherent imaging system. For fine resolution SAR images speckle can obscure subtle features and reduce visual appeal. Many speckle reduction methods result in a loss of image resolution and reduce visual appeal which can obscure subtle features. Another approach to maintain resolution while reducing speckle is to register and combine multiple images. For persistent surveillance applications it is more efficient for an airborne platform to fly circles around the particular area of interest. In these cases, it would be beneficial to combine multiple circle mode SAR images, however the image registration process is not so straightforward because the layover angle changes in each image. This paper develops a SAR image registration process for combining multiple circle mode SAR images to reduce speckle while preserving resolution. The registration first uses a feature matching algorithm for a coarse rotation and alignment, and then uses a fine registration and warp. Ku band SAR data from a circle mode SAR collection is used to show the effectiveness of the registration and enhanced visual appeal from multi-looking.

For synthetic aperture radars radio frequency interference from sources external to the radar system and techniques to mitigate the interference can degrade the quality of the image products. Usually the radar system designer will try to balance the amount of mitigation for an acceptable amount of interference to optimize the image quality. This dissertation examines the effect of interference mitigation upon coherent data products of fine resolution, high frequency synthetic aperture radars using stretch processing. Novel interference mitigation techniques are introduced that operate on single or multiple apertures of data that increase average coherence compared to existing techniques. New metrics are applied to evaluate multiple mitigation techniques for image quality and average coherence. The underlying mechanism for interference mitigation techniques that affect coherence is revealed.

Introduced a novel contrast metric in combination with traditional metrics to evaluate quality of coherent aperture radar, data products.

Abstract not provided.

Interference and interference mitigation techniques degrade synthetic aperture radar (SAR) coherent data products. Radars utilizing stretch processing present a unique challenge for many mitigation techniques because the interference signal itself is modified through stretch processing from its original signal characteristics. Many sources of interference, including constant tones, are only present within the fast-time sample data for a limited number of samples, depending on the radar and interference bandwidth. Adaptive filtering algorithms to estimate and remove the interference signal that rely upon assuming stationary interference signal characteristics can be ineffective. An effective mitigation method, called notching, forces the value of the data samples containing interference to zero. However, as the number of data samples set to zero increases, image distortion and loss of resolution degrade both the image product and any second order image products. Techniques to repair image distortions,1 are effective for point-like targets. However, these techniques are not designed to model and repair distortions in SAR image terrain. Good terrain coherence is important for SAR second order image products because terrain occupies the majority of many scenes. For the case of coherent change detection it is the terrain coherence itself that determines the quality of the change detection image. This paper proposes an unique equalization technique that improves coherence over existing notching techniques. First, the proposed algorithm limits mitigation to only the samples containing interference, unlike adaptive filtering algorithms, so the remaining samples are not modified. Additionally, the mitigation adapts to changing interference power such that the resulting correction equalizes the power across the data samples. The result is reduced distortion and improved coherence for the terrain. SAR data demonstrates improved coherence from the proposed equalization correction over existing notching methods for chirped interference sources.

For synthetic aperture radar image products interference can degrade the quality of the images while techniques to mitigate the interference also reduce the image quality. Usually the radar system designer will try to balance the amount of mitigation for the amount of interference to optimize the image quality. This may work well for many situations, but coherent data products derived from the image products are more sensitive than the human eye to distortions caused by interference and mitigation of interference. This dissertation examines the e ect that interference and mitigation of interference has upon coherent data products. An improvement to the standard notch mitigation is introduced, called the equalization notch. Other methods are suggested to mitigation interference while improving the quality of coherent data products over existing methods.

We propose a new laboratory method for characterizing synthetic aperture radar (SAR) systems through the use of a synthetic scene generator. Flight tests are the only definitive way to characterize the system level performance of airborne synthetic aperture radar systems. However, due to the expense of flights tests it is beneficial to complete as much testing as possible in a laboratory environment before flight testing is performed. There are many existing tests that are employed to measure the performance of various subsystems in a SAR system, find defective hardware, and indicate design problems that need to be mitigated. However, certain issues can only be found on an integrated system, and laboratory testing at a system level is typically confined to characterizing the impulse response (IPR) of a single point target through the use of an optical delay line. While useful, delay line testing requires running a modified version of real-time image formation code as the delay line does not completely mimic a real target. Ideally, system level tests are performed on unmodified code. On modern SAR systems many algorithms are data driven (e.g., autofocus) and require a substantially more sophisticated data model for testing. We desire to create a complete system test by combining an arbitrary number of point targets and clutter patterns to mimic radar responses from a real scene. This capability enables complete testing of radar systems in a laboratory environment according to prescribed terrain/scene characteristics. This paper presents an overview of the system requirements for a synthetic scene generator. The analysis is limited to SAR systems utilizing chirp waveforms and stretch processing. Furthermore, we derive relationships between IF bandwidth, target position, and the phase history model. A technique to properly compensate for motion pulse to pulse is presented. Finally, our concept is demonstrated with simulation data. © 2014 SPIE.

We propose a new laboratory method for characterizing synthetic aperture radar (SAR) systems through the use of a synthetic scene generator. Flight tests are the only definitive way to characterize the system level performance of airborne synthetic aperture radar systems. However, due to the expense of flights tests it is beneficial to complete as much testing as possible in a laboratory environment before flight testing is performed. There are many existing tests that are employed to measure the performance of various subsystems in a SAR system, find defective hardware, and indicate design problems that need to be mitigated. However, certain issues can only be found on an integrated system, and laboratory testing at a system level is typically confined to characterizing the impulse response (IPR) of a single point target through the use of an optical delay line. While useful, delay line testing requires running a modified version of real-time image formation code as the delay line does not completely mimic a real target. Ideally, system level tests are performed on unmodified code. On modern SAR systems many algorithms are data driven (e.g., autofocus) and require a substantially more sophisticated data model for testing. We desire to create a complete system test by combining an arbitrary number of point targets and clutter patterns to mimic radar responses from a real scene. This capability enables complete testing of radar systems in a laboratory environment according to prescribed terrain/scene characteristics. This paper presents an overview of the system requirements for a synthetic scene generator. The analysis is limited to SAR systems utilizing chirp waveforms and stretch processing. Furthermore, we derive relationships between IF bandwidth, target position, and the phase history model. A technique to properly compensate for motion pulse to pulse is presented. Finally, our concept is demonstrated with simulation data. © 2014 SPIE.

Synthetic aperture radar systems that use the polar format algorithm are subject to a focused scene size limit inherent to the polar format algorithm. The classic focused scene size limit is determined from the dominant residual range phase error term. Given the many sources of phase error in a synthetic aperture radar, a system designer is interested in how much phase error results from the assumptions made with the polar format algorithm. Autofocus algorithms have limits to the amount and type of phase error that can be corrected. Current methods correct only one or a few terms of the residual phase error. A system designer needs to be able to evaluate the contribution of the residual or uncorrected phase error terms to determine the new focused scene size limit. This paper describes a method to estimate the complete residual phase error, not just one or a few of the dominant residual terms. This method is demonstrated with polar format image formation, but is equally applicable to other image formation algorithms. A benefit for the system designer is that additional correction terms can be added or deleted from the analysis as necessary to evaluate the resulting effect upon image quality. © 2013 SPIE.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Since the Polar Format Algorithm (PFA) was first introduced by Jack Walker 30 years ago, digital processing and Moore’s law have provided the means by which to process an increasing amount of data, at finer resolutions, over a larger area, and in real-time. Inherent in the polar format algorithm are assumptions that limit the focused scene size. This report presents a development of PFA for a linear frequency modulated chirp pulsed radar utilizing stretch processing to illustrate how PFA approximations are used to form an image. Also techniques to mitigate the errors resulting from the approximations are presented from a survey of literature sources. There are many techniques that are successful at increasing the focused scene size, these include image corrections made after image formation, subaperture processing, and careful selection of processing coordinates. This report only considers methods that use the polar format algorithm.

Abstract not provided.

Abstract not provided.

Abstract not provided.