## THROUGH WALL IMAGING WITH UWB RADAR SYSTEM (2009)

### BibTeX

@MISC{Aftanas09throughwall,

author = {Ing Michal Aftanas},

title = {THROUGH WALL IMAGING WITH UWB RADAR SYSTEM},

year = {2009}

}

### OpenURL

### Abstract

Supervisor: doc. Ing. Miloˇs Drutarovsk´y, CSc. ”If we save even one life, we have been cost effective.”

### Citations

3098 | A computational approach to edge detection
- Canny
- 1986
(Show Context)
Citation Context ...plied. The peaks in Hough space are filtered according to the wall direction assumptions and lines which represent the investigated walls are drawn back to the Iw(xT , zT ) image. Canny edge detector =-=[25]-=- is used to detect edges, as it is very often used for detection of lines in combination with HT in SAR images obtained by an aircraft [33] or for ground penetrating radar SAR scans [56]. The HT is us... |

1339 |
Classical Electrodynamics
- Jackson
- 1999
(Show Context)
Citation Context ...jects are made up of perfect conductors, whereas objects made up of strong conductors are almost impenetrable in the sense that the impinging wave penetrates inside the object only to a shallow depth =-=[75]-=-. For such a class of objects it is thus possible to reconstruct only the object shapes but not to ”look” inside them. This is what happens at optical frequencies, when the impinging electromagnetic r... |

553 |
Use of the Hough transformation to detect lines and curves in pictures
- Duda, Hart
- 1972
(Show Context)
Citation Context ...ocessing steps: The image Iw(xT , zT ) is transformed to the gray-level. The edges are found with the edge detector so the picture is converted to the binary image IwDE(xT , zT ). The Hough transform =-=[129, 52]-=- is applied. The peaks in Hough space are filtered according to the wall direction assumptions and lines which represent the investigated walls are drawn back to the Iw(xT , zT ) image. Canny edge det... |

456 |
Multiple emitter location and signal parameter estimation
- Schmidt
- 1986
(Show Context)
Citation Context ...ain at a time corresponding to the arrival of the first reflection. An alternative method of resolving overlapping echoes is based on the use of the MUltiple SIgnal Characterization (MUSIC) algorithm =-=[127, 32]-=-. It is a high resolution spectral estimation method and is used to estimate the received signals covariance and then perform a spectral decomposition. 2.5 Radar Imaging Methods Overview SAR scanning ... |

420 |
Regularization of Inverse Problems
- Engl, Hank, et al.
- 1996
(Show Context)
Citation Context ...e whole radar system impulse response to the measured data is the most complex calibration process. There exist lots of more or less complicated methods how to provide deconvolution described e.g. in =-=[125,123,42,126,113,11,55]-=-. Long impulse response of antenna causes that any object that reflects the wave energy and is received by RX is presented in data also in long area. Influence of the whole radar system, and mainly th... |

296 |
M: Principles of Computerized Tomographic Imaging
- Kak, Slaney
- 1999
(Show Context)
Citation Context ... ideal approach and hence produces a lot of artifacts [96]. Conflict in names of basic imaging methods becomes evident in simple geometrical approach. This method is often called also back projection =-=[140,71,78,47,87]-=- or diffraction summation [98,112,96,108,99]. Moreover, this geometrical approach is often incorrectly called Kirchhoff migration [46, 35, 34, 153, 154]. 2.5.3 Simple Wall Compensation In through-wall... |

195 |
Computer Vision
- Shapiro, Stockman
- 2000
(Show Context)
Citation Context ...ocessing steps: The image Iw(xT , zT ) is transformed to the gray-level. The edges are found with the edge detector so the picture is converted to the binary image IwDE(xT , zT ). The Hough transform =-=[129, 52]-=- is applied. The peaks in Hough space are filtered according to the wall direction assumptions and lines which represent the investigated walls are drawn back to the Iw(xT , zT ) image. Canny edge det... |

145 |
Fast Fourier transforms for nonequispaced data
- Dutt, Rokhlin
- 1993
(Show Context)
Citation Context ...n airborne, or GPR scenarios. The most important of them will be shortly described in this section.2.5 Radar Imaging Methods Overview 30 Nonuniform FFT NonUniform Fast Fourier Transformation (NUFFT) =-=[53,19,91]-=- is based on classical Stolt migration for non equally spaced data. Stolt migration requires equally spaced input data for FFT algorithms. In many practical situations, however, the input data is nonu... |

105 |
Linograms in image reconstruction from projections
- Edholm
- 1987
(Show Context)
Citation Context ... ideal approach and hence produces a lot of artifacts [96]. Conflict in names of basic imaging methods becomes evident in simple geometrical approach. This method is often called also back projection =-=[140,71,78,47,87]-=- or diffraction summation [98,112,96,108,99]. Moreover, this geometrical approach is often incorrectly called Kirchhoff migration [46, 35, 34, 153, 154]. 2.5.3 Simple Wall Compensation In through-wall... |

95 |
On the fast Fourier transform of functions with singularities
- Beylkin
- 1995
(Show Context)
Citation Context ...n airborne, or GPR scenarios. The most important of them will be shortly described in this section.2.5 Radar Imaging Methods Overview 30 Nonuniform FFT NonUniform Fast Fourier Transformation (NUFFT) =-=[53,19,91]-=- is based on classical Stolt migration for non equally spaced data. Stolt migration requires equally spaced input data for FFT algorithms. In many practical situations, however, the input data is nonu... |

72 |
Introduction to Inverse Problems
- Bertero, Boccacci
- 1998
(Show Context)
Citation Context ...rse problem and in the last time this topic becomes an individual science discipline. In conjunction with radar imaging some of the good works can be found in [42] - chapter 7.9, [79] - chapter 6, or =-=[104,12,18,20,37]-=-. In this work only a basic principle is introduced. From the inverse problem point of view, migration can be taken as an approximate solution to the general elastic wavefield inversion problem [63]. ... |

70 |
Migration by Fourier Transform
- Stolt
(Show Context)
Citation Context ...vantages of this method. 2.5.6 f-k Migration Wave equation based migration could be done also in the frequency domain. Stolt in 1978 showed that migration problem could be solved by Fourier transform =-=[133]-=-.2.5 Radar Imaging Methods Overview 25 The basic principle for 2D problem is described. Complete mathematical solution is derived in [133] or in [96], therefore only the basic relations and principle... |

54 |
Earth soundings analysis: Processing versus inversion
- Claerbout
- 1992
(Show Context)
Citation Context ...ch. This method is often called also back projection [140,71,78,47,87] or diffraction summation [98,112,96,108,99]. Moreover, this geometrical approach is often incorrectly called Kirchhoff migration =-=[46, 35, 34, 153, 154]-=-. 2.5.3 Simple Wall Compensation In through-wall imaging, the electromagnetic wave has to penetrate through the wall, reflect from the investigated object, and penetrate back through the wall to the r... |

41 |
Super-resolution processing of multi-static data using time reversal and MUSIC,” Northeastern University Report, available at http://www.ece.neu.edu/faculty/devaney/ajd/preprints.htm
- Devaney
(Show Context)
Citation Context ...s and was introduced by Therrien in 1992 [138]. It is the method for estimating individual frequencies of multiple harmonic signals. The use of MUSIC for radar imaging was proposed by Devaney in 2000 =-=[48]-=-. He applied the algorithm to the problem of estimating the locations of a number of point-like scatterers. Detailed description of MUSIC algorithm can be found e.g. in [48, 32, 82]. MUSIC for image p... |

33 |
The MUSIC algorithm and the factorization method in inverse scattering theory for inhomogeneous media, Inverse Problems
- Kirsch
(Show Context)
Citation Context ...roposed by Devaney in 2000 [48]. He applied the algorithm to the problem of estimating the locations of a number of point-like scatterers. Detailed description of MUSIC algorithm can be found e.g. in =-=[48, 32, 82]-=-. MUSIC for image processing has been developed for multistatic radar systems. It is the wave equation based method. The basic idea of MUSIC is the formulation of the so called multistatic response ma... |

33 |
A new slant on seismic imaging: migration and integral geometry,” Geoph
- Miller, Oristaglio, et al.
- 1987
(Show Context)
Citation Context ... artifacts [96]. Conflict in names of basic imaging methods becomes evident in simple geometrical approach. This method is often called also back projection [140,71,78,47,87] or diffraction summation =-=[98,112,96,108,99]-=-. Moreover, this geometrical approach is often incorrectly called Kirchhoff migration [46, 35, 34, 153, 154]. 2.5.3 Simple Wall Compensation In through-wall imaging, the electromagnetic wave has to pe... |

33 |
The regular Fourier matrices and nonuniform fast Fourier transforms
- Nguyen, Liu
- 1999
(Show Context)
Citation Context ...s why the regular FFT can not be applied. NUFFT proposes an approach to achieve the fast Fourier transform for the nonuniform data by using the another class of matrices, the regular Fourier matrices =-=[105]-=-. This algorithm with the same complexity like Stolt migration is more accurate, because an approximation error is minimized in the leastsquare sense. Stolt migration requires so called Stolt interpol... |

32 |
A simple method using Morozov’s discrepancy principle for solving inverse scattering problems
- Colton, Piana, et al.
(Show Context)
Citation Context ...rse problem and in the last time this topic becomes an individual science discipline. In conjunction with radar imaging some of the good works can be found in [42] - chapter 7.9, [79] - chapter 6, or =-=[104,12,18,20,37]-=-. In this work only a basic principle is introduced. From the inverse problem point of view, migration can be taken as an approximate solution to the general elastic wavefield inversion problem [63]. ... |

31 |
Time-reversal imaging with multiple signal classification considering multiple scattering between the targets
- Gruber, Marengo, et al.
- 2004
(Show Context)
Citation Context ...ctrum is very large (theoretically, in the case of exact data, this value will be infinity). Simulated time-reversal imaging with MUSIC considering multiple scattering between the targets is shown in =-=[64]-=- In this case influence of a wall can be taken into account by applying Green’s function considering the wall, what is a real challenge itself.2.5 Radar Imaging Methods Overview 36 2.5.10 Singular Va... |

27 |
F.,1967: "Numerical Applications of a Formalism for Geophysical Inverse Problems
- Backus, Gilbert
(Show Context)
Citation Context ...rse problem and in the last time this topic becomes an individual science discipline. In conjunction with radar imaging some of the good works can be found in [42] - chapter 7.9, [79] - chapter 6, or =-=[104,12,18,20,37]-=-. In this work only a basic principle is introduced. From the inverse problem point of view, migration can be taken as an approximate solution to the general elastic wavefield inversion problem [63]. ... |

23 |
Integral formulation for migration in two and three dimensions
- Schneider
- 1978
(Show Context)
Citation Context ...ement that the integration surface must extend all the way around the volume containing the observation point must be dropped. There are various ways how to fulfill both of these arguments. Schneider =-=[128]-=- solved the requirement of ∂nψ by using a dipole Green’s function with an image source above the recording place, that vanished at z = 0 and cancelled the ∂nψ term in (2.5.22). Wiggins in [143] adapte... |

21 |
Geophysical signal analysis
- Robinson, Treitel
- 1980
(Show Context)
Citation Context ...hen used for deconvolution. There are several methods [42] like optimum or least squares filtering, Wiener filtering, matched filtering, minimatched filtering, etc. which performances are compared in =-=[112]-=-. The selection of a suitable filter depends on the characteristics of the signal. As the sample data set is composed of signal, noise and clutter, the question of the stability of the filter must be ... |

20 | The linear sampling method and the MUSIC algorithm
- Cheney
- 2001
(Show Context)
Citation Context ...ain at a time corresponding to the arrival of the first reflection. An alternative method of resolving overlapping echoes is based on the use of the MUltiple SIgnal Characterization (MUSIC) algorithm =-=[127, 32]-=-. It is a high resolution spectral estimation method and is used to estimate the received signals covariance and then perform a spectral decomposition. 2.5 Radar Imaging Methods Overview SAR scanning ... |

15 |
Extended chirp scaling algorithm for air- and spaceborne SAR data processing in stripmap and ScanSAR imaging modes
- Moreira, Mittermayer, et al.
- 1996
(Show Context)
Citation Context ...pensation for airborne measurements is described in [110]. Wavenumber-domain processing is limited because of its inability to perform high-precision motion compensation. Extended Chirp Scaling (ECS) =-=[100]-=- algorithm has proven to be very powerful, although it has certain limitations concerning long aperture syntheses and highly squinted geometries. Analytically derived new stripmap SAR data processing ... |

15 |
Ultra-wideband through-the-wall propagation
- Muqaibel, Safaai-Jazi, et al.
- 2005
(Show Context)
Citation Context ...pagation loss inside the wall is a function of the frequency [147]. Electromagnetic waves are able to penetrate through the concrete walls without massive attenuation of up to approximately 3 - 4 GHz =-=[101]-=-. With regard to this fact, M-sequence UWB radar technology is very attractive for through-wall object imaging. TOA is commonly referred to the flight time of the wave between transmit antenna, target... |

14 |
Prestack residual migration in the frequency domain
- Sava
- 2003
(Show Context)
Citation Context ... obtained in terms of enhanced mainlobe, suppressed sidelobes, and reduced noise. Prestack Residual f-k Migration Prestack residual migration in the frequency domain was introduced by Paul C. Sava in =-=[122]-=-. This method has advantages over classical f-k migration that estimates interval velocity functions for depth migration. It is more accurate than methods which are based on focusing the stack of migr... |

13 |
Convolution backprojection image reconstruction for spotlight mode synthetic aperture radar
- Desai, K
- 1992
(Show Context)
Citation Context ... ideal approach and hence produces a lot of artifacts [96]. Conflict in names of basic imaging methods becomes evident in simple geometrical approach. This method is often called also back projection =-=[140,71,78,47,87]-=- or diffraction summation [98,112,96,108,99]. Moreover, this geometrical approach is often incorrectly called Kirchhoff migration [46, 35, 34, 153, 154]. 2.5.3 Simple Wall Compensation In through-wall... |

13 |
An accurate algorithm for nonuniform fast Fourier transforms Ž NUFFT’s
- Liu, Nguyen
(Show Context)
Citation Context ...n airborne, or GPR scenarios. The most important of them will be shortly described in this section.2.5 Radar Imaging Methods Overview 30 Nonuniform FFT NonUniform Fast Fourier Transformation (NUFFT) =-=[53,19,91]-=- is based on classical Stolt migration for non equally spaced data. Stolt migration requires equally spaced input data for FFT algorithms. In many practical situations, however, the input data is nonu... |

13 |
Through-Wall Imaging (TWI) by Radar: 2-D Tomographic Results and Analysis
- Song, Yu, et al.
- 2005
(Show Context)
Citation Context ... can be seen that results obtained by inverse problem are better than by the conventional Kirchhoff migration. The simulations that use inverse problem in through-wall imaging field are shown e.g. in =-=[157, 132, 131]-=-. This technique promises reducing effect of multiple reflections between investigated objects including wall, as well as precise position, shape, and material characterization estimation. However, on... |

12 | Image estimation by example: geophysical soundings image construction
- Claerbout, Fomel
- 2010
(Show Context)
Citation Context ...ch. This method is often called also back projection [140,71,78,47,87] or diffraction summation [98,112,96,108,99]. Moreover, this geometrical approach is often incorrectly called Kirchhoff migration =-=[46, 35, 34, 153, 154]-=-. 2.5.3 Simple Wall Compensation In through-wall imaging, the electromagnetic wave has to penetrate through the wall, reflect from the investigated object, and penetrate back through the wall to the r... |

10 |
Synthetic Aperture Beamformer for Imaging Through a Dielectric Wall
- AHMAD, AMIN, et al.
- 2005
(Show Context)
Citation Context ...on of target shape and position [141]. Therefore, more or less accurate methods based on a ray theory and Snell’s law are used in algorithms of T OA computation to compensate the presence of the wall =-=[61, 5, 86, 6]-=-. The T OA for multilayer model cannot be computed directly, some numerical minimization method [106], has to be chosen. In general, the optimization techniques use T OA as an error function, and a ve... |

10 | Refocusing through building walls using synthetic aperture radar - Dehmollaian, Sarabandi |

10 |
Theory of Seismic Imaging
- Scales
- 1997
(Show Context)
Citation Context ...= dsurf = vr ∂t S0 r/v S0 4 cos θ vr [ ] ∂ψ dsurf. (2.5.28) ∂t 2r/v Finally, (2.5.28) is the Kirchhoff migration equation. This result was derived by many authors including Schneider [128] and Scales =-=[124]-=-. It expresses migration by summation along hyperbolic travel paths through the input data space. The hyperbolic summation does not have to be seen at the first view, but it can be indicated by [∂tψ] ... |

9 |
Ultra wideband radar systems: advantages and disadvantages
- Immoreeve, Fedotov
- 2002
(Show Context)
Citation Context ...ransmitted signal Time delay Received signal time time Fig. 2.1.1: Basic principle of the UWB radar. There are many advantages why to use UWB radar instead of radars with continuos waves described in =-=[74, 109]-=-. To name just a few most important the UWB signal can be transmitted with no carrier, the produced transmit signal requires less power, improved range measurement accuracy and object identification (... |

9 |
Through-wall imaging via a linear inverse scattering algorithm
- Soldovieri, Solimene
- 2007
(Show Context)
Citation Context ... can be seen that results obtained by inverse problem are better than by the conventional Kirchhoff migration. The simulations that use inverse problem in through-wall imaging field are shown e.g. in =-=[157, 132, 131]-=-. This technique promises reducing effect of multiple reflections between investigated objects including wall, as well as precise position, shape, and material characterization estimation. However, on... |

9 |
Imaging Through Unknown Walls Using Different Standoff Distances
- Wang, Amin
- 2006
(Show Context)
Citation Context ...ethod of the TOA for through-wall model is presented, published by the author in [1]. The conventional method that uses constant velocity model produces errors in object shape and position estimation =-=[141,4,3]-=-. Computation of the TOA (corresponding to the true flight distance) for three layer model requires the complex minimization algorithm. The proposed method transforms three layer (air-wallair) model t... |

8 |
Weakly closed nonlinear operators and parameter identification in parabolic equations by Tikhonov regularization
- Binder, Engl, et al.
- 1994
(Show Context)
Citation Context |

8 |
A brief comparison of some Kirchhoff integral formulas for migration and inversion
- Docherty
- 1991
(Show Context)
Citation Context ...function with an image source above the recording place, that vanished at z = 0 and cancelled the ∂nψ term in (2.5.22). Wiggins in [143] adapted Schneider’s technique to rough topography. Docherty in =-=[51]-=- showed that a monopole Green’s can also leads to an acceptable result and once again challenged Schneider’s argument, so the integral over the infinite hemisphere could be neglected. After all, migra... |

8 |
Surveillance Through Concrete Walls
- Gauthier, Chamma
- 2004
(Show Context)
Citation Context ...on of target shape and position [141]. Therefore, more or less accurate methods based on a ray theory and Snell’s law are used in algorithms of T OA computation to compensate the presence of the wall =-=[61, 5, 86, 6]-=-. The T OA for multilayer model cannot be computed directly, some numerical minimization method [106], has to be chosen. In general, the optimization techniques use T OA as an error function, and a ve... |

8 |
Three-dimensional Ground Penetrating Radar Imaging Using Multi-frequency Diffraction Tomography
- Mast, Johansson
- 1994
(Show Context)
Citation Context ...eat precision because at the final stage it would significantly affects the results. The radar imaging is well known field for a long time, mostly used for Ground Penetrating Radar (GPR) applications =-=[76]-=-. However, through-wall imaging has become investigated only a few years ago [60]. The most of the published methods were tested only on simulated data, and results from practical measurements are sho... |

8 |
Interpolating Cubic Splines
- Knott
- 1999
(Show Context)
Citation Context ...bration and preprocessing methods are applied individually to the A-scans. 2.4.1 Time Domain Interpolation The measured B-scan B(X, n) (Fig. 2.4.1 a)) is oversampled by the cubic spline interpolation =-=[83]-=- with an oversampling factor KO in the discrete time domain, so2.4 Calibration and Preprocessing 11 Amplitude 0.5 0 a) Measured A-scan Wall Object behind the wall Crosstalk Amplitude -1 0.5 0 5 10 15... |

8 |
The wave equation applied to migration: Geophys
- Loewenthal, Lu, et al.
- 1976
(Show Context)
Citation Context ...to estimate reflectivity, not just the scattered wavefield, and for this purpose a model relating the wavefield to the reflectivity is required. The most simple model is the exploding reflector model =-=[93]-=- which asserts that the reflectivity is identical to the downward continued scattered wavefield at t = 0 provided that the downward continuation is done with v = v/2. Thus, a wave migration equation f... |

8 |
Time domain modeling of UWB GPR and its application on landmine detection
- Scheers, Acheroy, et al.
- 2000
(Show Context)
Citation Context ...e whole radar system impulse response to the measured data is the most complex calibration process. There exist lots of more or less complicated methods how to provide deconvolution described e.g. in =-=[125,123,42,126,113,11,55]-=-. Long impulse response of antenna causes that any object that reflects the wave energy and is received by RX is presented in data also in long area. Influence of the whole radar system, and mainly th... |

7 |
Wave fi eld extrapolation techniques in seism.ic l11_igration , a tutorial
- Berkhout
- 1981
(Show Context)
Citation Context ...ds are based on a geometrical approach. After the introduction of computers, more complex techniques based on the scalar wave equation were introduced. A good overview of these techniques is given in =-=[150,17]-=-. 2.5.2 SAR Imaging - Geometrical Migration In this Section a conventional SAR imaging with geometrically based bistatic migration will be described. All geometrically based imaging methods perform on... |

7 |
Time reversal imaging for sensor networks with optimal compensation in time
- Derveaux, Papanicolaou, et al.
(Show Context)
Citation Context ...ch. This method is often called also back projection [140,71,78,47,87] or diffraction summation [98,112,96,108,99]. Moreover, this geometrical approach is often incorrectly called Kirchhoff migration =-=[46, 35, 34, 153, 154]-=-. 2.5.3 Simple Wall Compensation In through-wall imaging, the electromagnetic wave has to penetrate through the wall, reflect from the investigated object, and penetrate back through the wall to the r... |

7 |
Introduction to Global Optimization, 2 nd Ed
- Horst, Pardalos, et al.
- 2000
(Show Context)
Citation Context ...l’s law are used in algorithms of T OA computation to compensate the presence of the wall [61, 5, 86, 6]. The T OA for multilayer model cannot be computed directly, some numerical minimization method =-=[106]-=-, has to be chosen. In general, the optimization techniques use T OA as an error function, and a vector of independent variables over which the error is minimized [76]. The number of variables is equa... |

7 |
Extended wavenumber-domain synthetic aperture radar focusing with integrated motion compensation
- Reigber, Alivizatos, et al.
- 2006
(Show Context)
Citation Context ...ated by the operating wavelength and the transducer characteristics. f-k Migration with Motion Compensation f-k migration with integrated motion compensation for airborne measurements is described in =-=[110]-=-. Wavenumber-domain processing is limited because of its inability to perform high-precision motion compensation. Extended Chirp Scaling (ECS) [100] algorithm has proven to be very powerful, although ... |

7 |
Synthetic aperture radar processing using fast factorized back-projection
- Ulander, Hellsen, et al.
- 2003
(Show Context)
Citation Context |

6 |
High-Resolution Imaging using Capon Beamformers for Urban Sensing Applications
- Ahmad, Amin
- 2007
(Show Context)
Citation Context ...on of target shape and position [141]. Therefore, more or less accurate methods based on a ray theory and Snell’s law are used in algorithms of T OA computation to compensate the presence of the wall =-=[61, 5, 86, 6]-=-. The T OA for multilayer model cannot be computed directly, some numerical minimization method [106], has to be chosen. In general, the optimization techniques use T OA as an error function, and a ve... |

6 |
Autofocusing of through-the-wall radar imagery under unknown wall characteristics
- Ahmad, Amin, et al.
(Show Context)
Citation Context ...d incorrectly if the position of this object is estimated even with very little inaccuracy. Methods [141,142] were tested only on simulated data. SAR image de-smearing or auto-focusing were also used =-=[8, 95]-=-. However, the blurring of SAR images is practically very small and mostly lost in noise. Most of all the position of an object is changing significantly when the wall parameters are unknown as it has... |