Search

In this paper, we develop an optimal perturbation approach to design companders to mitigate the peak-to-average power ratio problem experienced by orthogonal frequency division multiplexing (OFDM) signals. Specifically, we generalize single-linear component design approaches, which modify the Rayleigh-distributed OFDM signal amplitude tail, to use of multiple contiguous piecewise linear components. The new compander design approach defines and solves a constrained optimization problem to optimally perturb a given proposed multicomponent piecewise linear segment set to meet the power and probability density function constraints. The use of multiple piecewise linear components provides greater design flexibility to meet the constraints, thus providing solutions for pairs of inflection points and cutoff values where more rigidly designed companders fail to exist. We present examples of compander design, and demonstrate compander performance through numerical simulation. The new compander design approach can generate companders that offer performance improvements over current companders.

Navigation of modern aerial platforms typically uses GPS position measurements for accurate position estimation. GPS can provide position accuracies of better than ten meters. However, operation in GPS-denied regions, precludes the use of GPS for navigation, thereby requiring other aiding sources for position estimation. To obtain altitude measurements, barometric-based altimeters are often used. Barometric-based altimeters provide a passive sensor alternative for low-cost aerial vehicles. These altimeters measure barometric pressure to estimate altitude, but provide errors of tens of meters or more due to atmospheric variation, weather conditions, and platform distance from altimeter calibration location. In this paper, we present a multi-camera vision based approach to altitude estimation, which can be used in the absence of GPS. This vision-based approach converts the altitude estimation problem to a ground-based image registration problem, for which highly accurate (pixel and sub-pixel-level accuracy) registration techniques are available. The altitude estimate results from simultaneously aligning multiple onboard cameras to ground reference imagery. The registration spatial error statistics corresponds directly to the altitude estimation error. We use image fusion over multiple registration solutions to provide accurate image registration and thus accurate altitude estimation. We first provide a technical description of the approach. We then generate numerical simulation results demonstrating performance of the approach, and validating aspects of the theory. We compare performance to a standard image rescaling approach for altitude estimation. We demonstrate improved performance over the classical approach.

The rapid expansion of data generated from various sensors, devices, and applications offers the opportunity to exploit complex data relationships in new ways. Multi-dimensional data often appears as multiple, separate data channels, for which multi-channel data representations and analysis techniques have been developed. Alternately, for image matching, multiple-template image matching techniques have been developed. Multi-template approaches use multiple, often singlechannel, templates, exhibiting intra-class variations. Same-class test image exemplars must match all reference templates. In this paper, we combine multiple-template matching techniques with multi-channel data representations to provide multitemplate, multi-channel image matching. We represent image data with pixels taking values in tensor products of lowdimensional Clifford Algebras, for which Fourier transforms exist. Fourier domain matching provides a computational processing improvement over spatial correlation-based matchers. The tensor product approach provides a decomposition of higher-dimensional algebras into combinations of lower dimensional algebras for which Fourier transforms apply. The tensor product approach produces a performance advantage, on data with the appropriate inter-channel correlation characteristics, through exploitation of additional data channel correlations. When these add constructively, a matching performance benefit occurs. We define an anti-involution mapping on a tensor product space, which leads to a definition of image correlation over these spaces. We prove that the correlation satisfies an extended inner product definition. We prove a Cauchy-Schwartz inequality, which validates use of the image correlation as a matcher. We present an example, using synthetic image data, where the approach provides superior matching performance over classical score sum fusion.

Effective infantry training requires accurate pointing direction knowledge of soldier-carried weapons to determine hit or miss scores. Infantry training solutions, which transition seamlessly to operational use, enables soldiers to train with what they use in combat, thereby increasing training efficacy and soldier lethality. To meet these dual needs, BAE Systems is developing the High-precision Infantry Training System (HITS). HITS provides a day/night, weapon-mounted ballistics point-of-impact detection system, to be used for live force-on-force training at Combat Training Centers (CTCs). HITS is hardware agnostic, but presently leverages BAE Systems’ fielded FWSI (Family-of-Weapons Sights-Individual) IR camera optic system to provide a ruggedized, low-SWaP, weapon-mounted platform for high-precision gun barrel pointing angle determination. The magnetometer and Inertial Measurement Unit (IMU) sensors, housed in the FWS-I gunsight, provide approximate pointing direction information while the FWS-I IR camera produces IR imagery with a field-of-view centered on the pointing direction of the gun. HITS uses innovative computer-vision processing to calculate the gun’s correct pointing angles. HITS performs geometric alignment of the FWS-I acquired IR image to a reference image of the view. The reference image is created by projecting a pre-generated reference model of the training site, consisting of geo-located 3-D LIDAR point-cloud data, onto the image plane. Using the gunsight location information, obtained from a GPS-RTK, the correct pointing angles are calculated from the image alignment. The system has an initial pointing accuracy goal of two milliradians. In this paper, we provide an overview of the HITS algorithm components including reference image generation, image registration, and pointing angle determination.

Recently, quaternions have been applied to multi-channel image processing such as color image registration, edge detection, and image matching. The general approach consists of using up to four image components as separate components in a quaternion-valued image. Techniques from quaternion algebra and analysis, such as quaternion Fourier transforms and correlation, are exploited to implement the desired image processing operations. The motivation underlying use of quaternions involves the concept of holistic data processing. That is, information in separate channels is processed jointly, thereby leveraging correlations across channel. Furthermore, quaternion transforms, such as the Fourier and wavelet transforms, have been defined and can be applied to process quaternion-valued imagery. For example, analogues from classical signal processing, such as the convolution theorem, have been derived and can be used to reduce computational burden. Such quaternion-based image processing techniques can provide performance superior to classical approaches which process image components independently. However, under certain conditions, performance of quaternion-based methods fails to meet performance of independent channel processing methods. In this paper, we perform a theoretical analysis of the performance of quaternion image matching. We present conditions under which quaternionbased processing outperforms, or falls short of, independent channel processing methods. We produce numerical image matching results validating the theoretical performance analysis using synthetic data. We define and solve a linear system of equations to generate the synthetic data. Receiver Operating Characteristics curves are provided to demonstrate the improved performance of the quaternion image matching approach under the favorable conditions predicted by the theory.

In recent years, the utility of multi-channel image processing techniques has increased. Applications to image processing include color image registration, edge detection, filtering, and color space correction, and multi-channel image matching. These techniques represent multiple image channels, such as three color channels, as image functions taking values in a finite-dimensional algebra, for example, the algebra of quaternions. Use of these algebras enables processing a multichannel image as a single entity, holistically. This holistic processing provides the opportunity to exploit cross-channel dependencies for improved performance, for example, improved matching performance in image matching or registration applications. In this paper, we make a critical analysis of the foundations of quaternion-valued image matching. We investigate standard alternatives to using quaternions for four-channel image matching including image tiling, image averaging, independent component processing, and coordinated component translation approaches. We examine the advantages that quaternion-based processing provides over these alternatives. We interpret the quaternion match metric as an inner product, which provides motivation for image matching using quaternion correlation. We generalize the quaternion match metric to use arbitrary combinations of component correlations to define new match metrics. We show that these new match metrics correspond to new product structures on four-dimensional algebras. We present numerical results demonstrating and validating aspects of the theory.

In this paper, we derive new multi-component companders for reducing peak-to-average power ratio (PAPR) in orthogonal frequency-division multiplexing signals, using a constrained optimization approach. The new companders provide out-of-band power rejection performance improvements over current state-of-the art companders, while maintaining comparable demodulation performance. The newly designed companders are derived by minimally perturbing the Rayleigh signal amplitude distribution, while enforcing the constant power and probability density function constraints. The use of a multi-component approach to compander construction provides design flexibility, and expands the space of tradeoffs between demodulation performance, PAPR reduction, and out-of-band power rejection. Furthermore, the new companders provide solutions in operating regions where certain current companders fail to exist due to violation of one or more constraints; for example, solutions may be derived for cutoff amplitude values that are unobtainable using other companders. We formulate the constrained optimization problem for compander design and derive the compander and decompander forms. Through numerical simulation, we generate performance results demonstrating the capability of the new companders.

While Machine Learning (ML) Automatic Target Recognition (ATR) represents the state-of-the art in target recognition, model-based ATR plays a valuable role. Model-based ATR complements machine learning ATR approaches by filling a near-term niche. While explainable Artificial Intelligence (AI) is not yet fully realized, model-based ATR serves to validate machine learning recognition decisions, and thus instills confidence in ML target calls. Alternatively, model-based ATR can act as a stand-alone ATR component, particularly in scenarios in which a small number of targets are of interest, e.g., “target-of-the-day” engagements. Model-based ATR approaches need no training data, and thus provide an alternative to machine learning approaches in the absence of sufficient quantities of real, or sufficiently high-fidelity synthetic, training data. In this paper, we present an approach to model-based ATR, called Shape-Based ATR (SB-ATR), which captures salient target shape information for recognizing targets in wide-area satellite imagery. SB-ATR finds the right blend of coarse 3-D target shape abstraction and target realism to provide robustness against target variations and environmental operating conditions, while simultaneously providing high-performance target recognition. The approach uses newer, robust forms of image correlation for matching a predicted target shape against the image. Shape prediction searches over target pose, and uses satellite metadata and solar geometry to generate realistic target shape and shadow predictions. The correlation matchers provide tolerance to illumination variations, moderate occlusions, image distortions and noise, and geometric differences between models and real targets. We present technical details of the shape-based approach, and provide numerical target recognition results on real-world satellite imagery demonstrating performance.

Accurate and successful image registration is a key enabling technology in applications such as image fusion, matching and pattern recognition. Knowledge of registration solution quality and accuracy can help prevent an inaccurate registration from degrading or corrupting performance of downstream image processing applications. However, estimating the spatial accuracy of image registration solutions can be difficult in the absence of ground-truth information on feature content or fiducial marker correspondences. This paper presents an automated spatial registration accuracy measurement for estimating and quantifying the spatial accuracy of correlation-based image registration in the absence of ground-truth information. For correlation surfaces exhibiting a single dominant peak, the approach consists of fitting an appropriate region of the correlation surface, about the peak coefficient, with a two-dimensional Gaussian. It then uses the covariance of the Gaussian to model the registration spatial error covariance. Use of a fitted Gaussian provides an intuitive probabilistic interpretation to the registration solution; the Gaussian function value at a spatial offset from the Gaussian peak gives the likelihood of that offset value. For more complicated regions containing multiple correlation local peak values, we extend the approach to fit a Gaussian mixture model to the region and use the mixture model covariance for the spatial accuracy metric. We describe an energy-based method for choosing the model region of the correlation surface. We discuss implementation subtleties and provide perturbation methods for handling numerically illconditioned matrices. We present numerical spatial error estimation results generated from registration of real video imagery acquired from a UAV platform.

The complexity of modern defense systems requires a holistic approach to modeling system performance. Reductionist methods fail to capture emergent behaviors and other complex behaviors characteristic of these systems. For sensor systems that include ground moving target indicator (GMTI) radar sensors, having a model for probability of detection is useful for simulating, analyzing, and evaluating system performance. To this end, this paper presents mathematical modeling of detection performance of a GMTI ground radar sensor system. Specifically, we define a detection scenario in which a single ground mover within the radar scan sector is randomly distributed on a road network according to traffic loading constraints. We perform mathematical modeling to generate a general truth expression for probability of ground mover detection over general road networks. We present special cases in which closed-form expressions for the probability of detection exist. We generate empirical detection results using Monte Carlo simulation with a GMTI radar simulation to validate the mathematical modeling.

The pervasiveness of small Unmanned Aerial Vehicles (UAVs), due to low cost, ease of control, and portability, opens the possibility of their use in urban environments for illegal or adversarial purposes. Such use includes unauthorized surveillance, reconnaissance, and weaponization. Detecting adversarial UAVs in urban environments is difficult. Urban canyons provide shielding from visibility. The small size of quadcopter-type UAVs limits the number of object pixels available for processing, which reduces standoff detection performance. UAVs fly against a background of ground motion clutter which can mask their motion. One possible solution to small UAV detection in urban environments uses low-cost UAV surveillance platforms, equipped with optical sensors, together with computer vision algorithms to detect adversarial UAVs in video data. In this paper we adapt the astronomical technique of transit photometry to detect small UAVs, operating in urban environments, in video data. Transit photometry, typically used for exo-planet discovery, detects small changes in background brightness due to a transiting object. As the UAV traverses across a bright background region, for example, the vehicle occludes the background and reduces the perceived brightness. This brightness dip may be used to infer the existence of a potential UAV passing across the background. The transit photometry curve, resulting from this brightness dip, reveals information about the traversing vehicle. We investigate mathematical properties of the transit photometry curve and derive a closed-form expression for it. We present numerical results demonstrating the technique on real video data acquired from a small UAV operating in an urban environment.

The ubiquity of mobile devices offers the opportunity to exploit device-generated signal data for biometric identification, health monitoring, and activity recognition. In particular, mobile devices contain an Inertial Measurement Unit (IMU) that produces acceleration and rotational rate information from the IMU accelerometers and gyros. These signals reflect motion properties of the human carrier. It is well-known that the complexity of bio-dynamical systems gives rise to chaotic dynamics. Knowledge of chaotic properties of these systems has shown utility, for example, in detecting abnormal medical conditions and neurological disorders. Chaotic dynamics has been found, in the lab, in bio-dynamical systems data such as electrocardiogram (heart), electroencephalogram (brain), and gait data. In this paper, we investigate the following question: can we detect chaotic dynamics in human gait as measured by IMU acceleration and gyro data from mobile phones? To detect chaotic dynamics, we perform recurrence analysis on real gyro and accelerometer signal data obtained from mobile devices. We apply the delay coordinate embedding approach from Takens' theorem to reconstruct the phase space trajectory of the multi-dimensional gait dynamical system. We use mutual information properties of the signal to estimate the appropriate delay value, and the false nearest neighbor approach to determine the phase space embedding dimension. We use a correlation dimension-based approach together with estimation of the largest Lyapunov exponent to make the chaotic dynamics detection decision. We investigate the ability to detect chaotic dynamics for the different one-dimensional IMU signals, across human subject and walking modes, and as a function of different phone locations on the human carrier.

An expression is derived for probability of correct association for maximum likelihood-based association using Gauss-Hermite quadrature. We derive a closed-form error bound characterizing accuracy. We generate empirical results using Monte Carlo simulation and show agreement with theoretical results. We show how two limit cases are obtainable from the integral expression.

In this paper, we characterize a family of closed-form companders for peak-to-average power ratio (PAPR) reduction in orthogonal frequency-division multiplexing (OFDM) signals. This new family generalizes the current state-of-the-art companders that are based on modification of the OFDM signal amplitude statistics. We derive the compander and decompander forms for a specific, parameterized member of this family. The larger parameter set characterizing these new companders provides greater design flexibility. The new companders are shown to provide solutions over regions of the operating conditions space where other current companders fail to exist. These new solutions thus widen the application of companders for PAPR reduction, and expand the trade-space over demodulation performance, PAPR reduction, and out-of-band power rejection for compander design. Through numerical result generation and simulation, we generate performance results to show that the new companders offer performance improvements over current companders.

Modern image processing performed on-board low Size, Weight, and Power (SWaP) platforms, must provide high- performance while simultaneously reducing memory footprint, power consumption, and computational complexity. Image preprocessing, along with downstream image exploitation algorithms such as object detection and recognition, and georegistration, place a heavy burden on power and processing resources. Image preprocessing often includes image denoising to improve data quality for downstream exploitation algorithms. High-performance image denoising is typically performed in the wavelet domain, where noise generally spreads and the wavelet transform compactly captures high information-bearing image characteristics. In this paper, we improve modeling fidelity of a previously-developed, computationally-efficient wavelet-based denoising algorithm. The modeling improvements enhance denoising performance without significantly increasing computational cost, thus making the approach suitable for low-SWAP platforms. Specifically, this paper presents modeling improvements to the Sendur-Selesnick model (SSM) which implements a bivariate wavelet shrinkage denoising algorithm that exploits interscale dependency between wavelet coefficients. We formulate optimization problems for parameters controlling deadzone size which leads to improved denoising performance. Two formulations are provided; one with a simple, closed form solution which we use for numerical result generation, and the second as an integral equation formulation involving elliptic integrals. We generate image denoising performance results over different image sets drawn from public domain imagery, and investigate the effect of wavelet filter tap length on denoising performance. We demonstrate denoising performance improvement when using the enhanced modeling over performance obtained with the baseline SSM model.

Navigation of airborne platforms in environments where access to the Global Positioning System is denied requires development of alternatives to GPS. Vision-aided navigation is an alternative that works by comparing feature content, from real-time data acquired by on-board vision sensors, to scene content in geo-referenced imagery. This approach generates 6-degree-of-freedom (6DOF) measurements for platform position and orientation. Development of high-performance vision-aided navigation solutions is challenging due to various effects including platform dynamics, data communication rates and sensor tasking. To meet these challenges, we are presenting the Vision-Aided Navigation Architecture (VANA). VANA serves as a vehicle for third-party developers to create applications (“apps”) under a plug-and-play applications paradigm. In addition to third-party apps, VANA includes a suite of default apps to provide baseline vision-aided navigation functionality. In VANA, multiple apps may run simultaneously using the same or different data sources. Additionally, VANA provide flexibility to use data from a variety of sensor modalities by abstracting sensors into three groups: global sensors (e.g., GPS), vehicle-referenced kinematic sensors (e.g., inertial measurement unit [IMU] and inertial navigation system [INS], and vision sensors (e.g., video, ranging). This grouping supports inclusion of additional sensors. All three sensor groups are tracked by a Global Navigation Manager (GNM), communicating with a Main Navigation Loop (MNL), which coordinates database access and data fusion apps. We also describe in detail the VANA Simulation Environment (VANA-SE) which enables rapid development, testing, and evaluation of apps. Our simulation environment provides a framework for component integration and testing, including application programming interfaces (API) for access by app developers. In this paper, we present the VANA architecture; functionality development details, including the complete VANA-SE framework; as well as initial simulation design, results and analysis, for the first spiral of VANA. Approved for public release; unlimited distribution.

Data fusion can be used to generate high quality data from multiple, degraded data sets by appropriately extracting and combining “good” information from each degraded set. In particular for image fusion, it may be used for image denoising, deblurring, or pixel dropout compensation. Image fusion is often performed in an image transform domain. In transform domain fusion approaches, transform coefficients from multiple images may be combined in various ways to produce an improved transform coefficient set. The fused transform data is inverted to produce the fused image. In this paper we formulate a general approach to image fusion in the wavelet domain. The proposed approach exploits context information, through application of nonparametric statistical hypothesis testing. The use of statistical hypothesis testing places the fusion on a theoretically sound and principled basis, and leads to improved fusion performance. Furthermore, use of statistical wavelet coefficient information in a neighborhood of the test coefficient more fully exploits the available context information. In this paper we first formulate the fusion approach. We then present numerical image data fusion results using a sampling of imagery from a public domain image database. We compare fusion performance of the proposed approach with performance of other standard wavelet-domain fusion approaches, and show a performance improvement when using the proposed approach.

Search algorithm complexity modeling, in the form of penetration rate estimation, provides a useful way to estimate search efficiency in application domains which involve searching over a hypothesis space of reference templates or models, as in model-based object recognition, automatic target recognition, and biometric recognition. The penetration rate quantifies the expected portion of the database that must be searched, and is useful for estimating search algorithm computational requirements. In this paper we perform mathematical modeling to derive general equations for penetration rate estimates that are applicable to a wide range of recognition problems. We extend previous penetration rate analyses to use more general probabilistic modeling assumptions. In particular we provide penetration rate equations within the framework of a model-based image alignment application domain in which a prioritized hierarchical grid search is used to rank subspace bins based on matching probability. We derive general equations, and provide special cases based on simplifying assumptions. We show how previously-derived penetration rate equations are special cases of the general formulation. We apply the analysis to model-based logo image alignment in which a hierarchical grid search is used over a geometric misalignment transform hypothesis space. We present numerical results validating the modeling assumptions and derived formulation.

Fusion techniques have proven to be very useful for many signal and image processing applications including image recognition, image registration, and biometric matching. Along with standard fusion techniques, hypercomplex image processing techniques have been developed recently. These techniques represent a form of image fusion in which several image components are combined to form a multi-channel image. The multi-channel imagery may be processed using hypercomplex transforms, such as the hypercomplex Fourier transform, for image matching and registration. In this paper we investigate performance of multi-channel image fusion for face image matching. We use 3-D color face imagery and investigate fusion of various combinations of grayscale intensity, color, and range information. We conduct a theoretical investigation to identify conditions under which matchers using image channel fusion provide superior matching performance relative to matchers fusing single channel image matching results. We present numerical performance results in the form of Receiver Operating Characteristics curves quantifying matching performance for verification hypothesis testing problems.

Feature extraction is a key component of typical pattern recognition algorithms. Usually performance of feature extractors is governed by several parameters. Characterizing parameter value effect on feature extraction performance is valuable for aiding in appropriate parameter value selection. Often, the parameter space is discretized and the effect of discrete parameter values on feature variation is analyzed. However, it can be problematic to determine a discretization density that contains suitable parameter values. To address this issue, this paper further explores a previously-introduced sensitivity measure for quantifying feature variation as a function of parameter space sampling density. Further mathematical properties of the sensitivity measure are determined. Closed form expressions for special feature set relationships are derived. We investigate sensitivity measure convergence properties as a function of increasing parameter space sampling density. We present conditions for sensitivity measure convergence, and provide closed form expressions for the limiting values. We show how sensitivity measure convergence can be used to choose an appropriate parameter space sampling density. Numerical examples of sensitivity measure convergence, validating the theoretical results, are presented for feature extraction on natural imagery.

Hypercomplex approaches are seeing increased application to signal and image processing problems. The use of multicomponent hypercomplex numbers, such as quaternions, enables the simultaneous co-processing of multiple signal or image components. This joint processing capability can provide improved exploitation of the information contained in the data, thereby leading to improved performance in detection and recognition problems. In this paper, we apply hypercomplex processing techniques to the logo image recognition problem. Specifically, we develop an image matcher by generalizing classical phase correlation to the biquaternion case. We further incorporate biquaternion Fourier domain alpha-rooting enhancement to create Alpha-Rooted Biquaternion Phase Correlation (ARBPC). We present the mathematical properties which justify use of ARBPC as an image matcher. We present numerical performance results of a logo verification problem using real-world logo data, demonstrating the performance improvement obtained using the hypercomplex approach. We compare results of the hypercomplex approach to standard multi-template matching approaches.

In computer vision applications, image matching performed on quality-degraded imagery is difficult due to image content distortion and noise effects. State-of-the art keypoint based matchers, such as SURF and SIFT, work very well on clean imagery. However, performance can degrade significantly in the presence of high noise and clutter levels. Noise and clutter cause the formation of false features which can degrade recognition performance. To address this problem, previously we developed an extension to the classical amplitude and phase correlation forms, which provides improved robustness and tolerance to image geometric misalignments and noise. This extension, called Alpha-Rooted Phase Correlation (ARPC), combines Fourier domain-based alpha-rooting enhancement with classical phase correlation. ARPC provides tunable parameters to control the alpha-rooting enhancement. These parameter values can be optimized to tradeoff between high narrow correlation peaks, and more robust wider, but smaller peaks. Previously, we applied ARPC in the radon transform domain for logo image recognition in the presence of rotational image misalignments. In this paper, we extend ARPC to incorporate quaternion Fourier transforms, thereby creating Alpha-Rooted Quaternion Phase Correlation (ARQPC). We apply ARQPC to the logo image recognition problem. We use ARQPC to perform multiple-reference logo template matching by representing multiple same-class reference templates as quaternion-valued images. We generate recognition performance results on publicly-available logo imagery, and compare recognition results to results generated from standard approaches. We show that small deviations in reference templates of sameclass logos can lead to improved recognition performance using the joint matching inherent in ARQPC.

Detecting dim targets in infrared imagery remains a challenging task. Several techniques exist for detecting bright, high contrast targets such as CFAR detectors, edge detection, and spatial thresholding. However, these approaches often fail for detection of targets with low contrast relative to background clutter. In this paper we exploit the transient capture capability and directional filtering aspect of wavelets to develop a wavelet based image enhancement method. We develop an image representation, using wavelet filtered imagery, which facilitates dim target detection. We further process the wavelet-enhanced imagery using the Michelson visibility operator to perform nonlinear contrast enhancement prior to target detection. We discuss the design of optimal wavelets for use in the image representation. We investigate the effect of wavelet choice on target detection performance, and design wavelets to optimize measures of visual information on the enhanced imagery. We present numerical results demonstrating the effectiveness of the approach for detection of dim targets in real infrared imagery. We compare target detection performance to performance obtained using standard techniques such as edge detection. We also compare performance to target detection performed on imagery enhanced by optimizing visual information measures in the spatial domain. We investigate the stability of the optimal wavelets and detection performance variation, across perspective changes, image frame sample (for frames extracted from infrared video sequences), and image scene content types. We show that the wavelet-based approach can usually detect the targets with fewer false-alarm regions than possible with standard approaches.

Realtime multisensor image registration algorithms must be computationally efficient. Often, simplifying assumptions are made to reduce computational time. However, these simplifications usually trade registration convergence performance for reduced runtime. For non-realtime applications where computational resources are not severely limited, this tradeoff may be reversed to improve convergence performance at the expense of increased computational cost. To this end we introduce a smart iterative approach to minimize mis-registrations and thus optimize registration convergence probability. The approach involves performing a registration sweep over a smart sampling of parameters governing feature generation. This approach involves use of two components; a feature sensitivity measure (FSM) and a registration verification metric (VM). The FSM measures the effect of parameter values on feature set variability. This measure enables choice of a suitable parameter sampling density to use for performing iterative registration solution search. The VM provides feedback on the registration solution verity in the absence of ground truth and is used to identify a converged solution. First, we provide an overview of the registration framework used to generate convergence results. Next we introduce the FSM and present mathematical properties. We then describe the VM and present the iterative algorithm. We present numerical results illustrating FSM convergence with increasing parameter sampling density for Canny edge features in SAR imagery. We illustrate use of FSM convergence behavior to select a suitable parameter sampling density for use in the iterative algorithm. Finally, SAR-to-EO registration performance results are presented showing improved convergence probability.

Spatial domain log-polar approaches have demonstrated success for video image registration. However, the log-polar representation is sensitive to origin location. This drawback often necessitates performing a parameter sweep over log-polar origin location, which can be time-consuming. In this paper we present an alternative approach that is appropriate for small to moderate scale, rotational, and skew misalignments but allows large translational offset. We use a form of robust phase correlation to estimate the gross translation, then perform a local search over log-polar origin to fine tune the translation. We sequentially estimate affine transform parameters by maximizing a measure of registration solution verity. We also investigate the effect of scale and rotational initial alignment errors on the robustness of the initial phase correlation to estimate gross translation. We present results using video imagery acquired from a real aerial video surveillance system.

This paper generalizes the previously developed automated edge-detection parameter selection algorithm of Yitzhaky and Peli. We generalize the approach to arbitrary multidimensional, continuous or discrete parameter spaces, and feature spaces. This generalization enables use of the parameter selection approach with more general image features, for use in feature-based multisensor image registration applications. We investigate the problem of selecting a suitable parameter space sampling density in the automated parameter selection algorithm. A real-valued sensitivity measure is developed which characterizes the effect of parameter space sampling on feature set variability. Closed-form solutions of the sensitivity measure for special feature set relationships are derived. We conduct an analysis of the convergence properties of the sensitivity measure as a function of increasing parameter space sampling density. For certain parameter space sampling sequence types, closed-form expressions for the sensitivity measure limit values are presented. We discuss an approach to parameter space sampling density selection which uses the sensitivity measure convergence behavior. We provide numerical results indicating the utility of the sensitivity measure for selecting suitable parameter values.

Persistent aerial video surveillance from small UAV (SUAV) platforms requires accurate and robust target tracking capabilities. However, video tracks can break due to excessive camera motion, target resolution, low signal-to noise ratio, video frame dropout, and frame-to-frame registration errors. Connecting broken tracks (video track repair) is thus essential for maintaining high quality target tracks. In this paper we present an approach to track repair based on multi-hypothesis sequential probability ratio tests (MHSPRT) that is suitable for real-time video tracking applications. To reduce computational complexity, the approach uses a target dynamics model whose state estimation covariance matrix has an analytic eigendecomposition. Chi-square gating is used to form feasible track-to-track associations, and a set of local hypothesis tests is defined for associating new tracks with coasted tracks. Evidence is accumulated across video frames by propagating posterior probabilities associated with each track repair hypothesis in the MHSPRT framework. Global maximum likelihood and maximum a posteriori estimation techniques resolve conflicts between local track association hypotheses. The approach also supports fusion of appearance-based features to augment statistical distributions of the track state and enhance performance during periods of kinematic ambiguity. First, an overview of the video tracker technology is presented. Next the track repair algorithm is described. Finally, numerical results are reported demonstrating performance on real video data acquired from an SUAV.

Small Unmanned Aerial Vehicles (UAVs) offer great potential to increase the war fighter’s short term situational awareness. Small UAVs are low cost and light weight, offering soldiers the ability to rapidly deploy a dedicated surveillance asset, thereby immediately increasing their tactical advantage over nearby adversaries. While this short look ahead is invaluable in modern tactical environments, the effectiveness of a small UAV is highly dependent on the operator’s ability to manage the vehicle’s flight path while interpreting the video surveillance data received. In this paper, we will discuss a video tracking approach to reduce operator workload. First, the operator designates a target or location on the ground to track. A video tracker then follows that designation by specifying waypoints and sensor look angles to keep that location in the center of the field of view.

Image registration is usually a required first processing step for such activities as surveillance, video tracking, change detection, and remote sensing. Often, different sensors are used for the collection of the test and reference imagery. The sensor phenomenology differences can present problems for automatic selection of registration algorithm parameters because of different cross-sensor feature manifestation. In previous work involving edge-based multisensor image registration, we applied a previously-developed automated approach to parameter selection, designed specifically for edge detection. In this work, we adapt and apply a dynamic feature selection algorithm (DFSA) that we recently developed for use in registration algorithm selection for registering images with varying scene content type. We adapt and apply the DFSA to the problem of selecting appropriate registration algorithm parameter values in an edge-based registration algorithm. The approach involves generating test-to-reference feature match scores over a sampling of the transform hypothesis space. The approach is scene-adaptive thereby requiring no a priori information on image scene content. Furthermore, in the DFSA we leverage prior match score calculation generated in a hierarchical grid search to reduce additional computational expense. We give a brief overview of the registration algorithmic framework, and present a description of the dynamic feature selection algorithm. Numerical results are presented for performing test SAR-to-reference EO image registration to show the registration convergence performance improvement resulting from use of the DFSA. Numerical results are generated over images exhibiting different scene content types. We also evaluate the effect of match score normalization on the registration convergence performance improvement.© (2007) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Accurate geo-location of imagery produced from airborne imaging sensors is a prerequisite for precision targeting and navigation. However, the geo-location metadata often has significant errors which can degrade the performance of applications using the imagery. When reference imagery is available, image registration can be performed as part of a bundle-adjustment procedure to reduce metadata errors. Knowledge of the metadata error statistics can be used to set the registration transform hypothesis search space size. In setting the search space size, a compromise is often made between computational expediency and search space coverage. It therefore becomes necessary to detect cases in which the true registration solution falls outside of the initial search space. To this end, we develop a registration verification metric, for use in a multisensor image registration algorithm, which measures the verity of the registration solution. The verification metric value is used in a hypothesis testing problem to make a decision regarding the suitability of the search space size. Based on the hypothesis test outcome, we close the loop on the verification metric in an iterative algorithm. We expand the search space as necessary, and re-execute the registration algorithm using the expanded search space. We first provide an overview of the registration algorithm, and then describe the verification metric. We generate numerical results of the verification metric hypothesis testing problem in the form of Receiver Operating Characteristics curves illustrating the accuracy of the approach. We also discuss normalization of the metric across scene content.

Accurate image registration is critical for applications such as precision targeting, geo-location, change-detection, surveillance, and remote sensing. However, the increasing volume of image data is exceeding the current capacity of human analysts to perform manual registration. This image data glut necessitates the development of automated approaches to image registration, including algorithm parameter value selection. Proper parameter value selection is crucial to the success of registration techniques. The appropriate algorithm parameters can be highly scene and sensor dependent. Therefore, robust algorithm parameter value selection approaches are a critical component of an end-to-end image registration algorithm. In previous work, we developed a general framework for multisensor image registration which includes feature-based registration approaches. In this work we examine the problem of automated parameter selection. We apply the automated parameter selection approach of Yitzhaky and Peli to select parameters for feature-based registration of multisensor image data. The approach consists of generating multiple feature-detected images by sweeping over parameter combinations and using these images to generate estimated ground truth. The feature-detected images are compared to the estimated ground truth images to generate ROC points associated with each parameter combination. We develop a strategy for selecting the optimal parameter set by choosing the parameter combination corresponding to the optimal ROC point. We present numerical results showing the effectiveness of the approach using registration of collected SAR data to reference EO data.

In this work we investigate simultaneous object identification improvement and efficient library search for model-based object recognition applications. We develop an algorithm to provide efficient, prioritized, hierarchical searching of the object model database. A common approach to model-based object recognition chooses the object label corresponding to the best match score. However, due to corrupting effects the best match score does not always correspond to the correct object model. To address this problem, we propose a search strategy which exploits information contained in a number of representative elements of the library to drill down to a small class with high probability of containing the object. We first optimally partition the library into a hierarchic taxonomy of disjoint classes. A small number of representative elements are used to characterize each object model class. At each hierarchy level, the observed object is matched against the representative elements of each class to generate score sets. A hypothesis testing problem, using a distribution-free statistical test, is defined on the score sets and used to choose the appropriate class for a prioritized search. We conduct a probabilistic analysis of the computational cost savings, and provide a formula measuring the computational advantage of the proposed approach. We generate numerical results using match scores derived from matching highly-detailed CAD models of civilian ground vehicles used in 3-D LADAR ATR. We present numerical results showing effects on classification performance of significance level and representative element number in the score set hypothesis testing problem.

LADAR imagery provides the capability to represent high resolution detail of 3D surface geometry of complex targets. In previous work we exploited this capability for automatic target recognition (ATR) by developing matching algorithms for performing surface matching of 3D LADAR point clouds with highly-detailed target CAD models. A central challenge in evaluating ATR performance is characterizing the degree of problem difficulty. One of the most important factors is the inherent similarity of target signatures. We've developed a flexible approach to target taxonomy based on 3D shape which includes a classification framework for defining the target recognition problem and evaluating ATR algorithm performance. The target model taxonomy consists of a hierarchical, tree-structured target classification scheme in which different levels of the tree correspond to different degrees of target classification difficulty. Each node in the tree corresponds to a collection of target models forming a target category. Target categories near the tree root represent large and very general target classes, exhibiting large interclass distance. Targets in these categories are easily separated. Target categories near the tree bottom represent very specific target classes with small interclass distance. These targets are difficult to separate. In this paper we focus on creation of optimal categories. We develop approaches for optimal aggregation of target model types into categories which provide for improved classification performance. We generate numerical results using match scores derived from matching highly-detailed CAD models of civilian ground vehicles.