Subsurface Target Recognition Based on Transient Electromagnetic Scattering

Successful extraction of the resonant modes embedded in the transient target signature is vital to resonance-based radar target recognition. For targets in free space, although these resonant modes are theoretically aspect independent, the corresponding residues are known to be aspect dependent. When the target is located above a halfspace, the interface will electromagnetically interact with the target and alter the residue. Within the context of resonance-based target recognition, the influence from an interface to the residues, to the best of our knowledge, has not been well studied. In this paper, we study this problem through a wire target. Our results show that the residues of the mode follow a similar behavior as the dielectric properties of the halfspace vary.

Radar target recognition is an important research direction in the field of radar information processing. Low resolution radar ground target classification and recognition still has broad application prospects in modern military and civil fields. Convolutional neural network (CNN) has attracted more and more attention in the field of radar automatic target recognition, because it does not need Feature Engineering and has superior classification performance. The optimization algorithm of neural network is proposed, and the effects of different network structures on the recognition effect are compared, in this paper. The results show that the multi-scale wide residual network multi-target classification algorithm based on optimized auto-encoder has better recognition effect in low resolution radar target recognition.

Radar scattering amplitudes contain pole singularities whose importance was recognized in the context of the Singularity Expansion Method: S.E.M. This method uses the fact that the late time domain response rt(t) of a target, illuminated by an E.M. wave, is mainly defined in a frequency band corresponding to the resonance region of the object. The knowledge of the singularities is useful information for discrimination of radar targets and has been used for different purposes of discrimination and identification. In this paper, we propose a modified scheme of radar target identification. The method presented is based on E-Pulse technique. In practical cases, direct application of classical E-Pulse techniques is not very efficient. Its performances are damaged by the characteristics of the exciting signal (antenna output signal). We propose a modified scheme of E-Pulse technique, which allows more accurate target discrimination and improves radar target identification. This procedure requires the deconvolution of the target response by the antenna signal and the application of an equivalent gaussian impulse excitation. This process has been successfully tested to FDTD simulations and measurements in anechoic chamber. 40 Toribio, Saillard, and Pouliguen

It is gaining acceptance by the radar signature community that effective target recognition requires more advanced signal analysis methods than those available for the frequency or time domains, separately. In recent years, the various distributions in the time-frequency domain of the Wigner-type, or Cohen class1–3 have proven their feasibility of extracting features in the cross-section of returned echoes and be more informative by their ability to display the time evolution of these features. As we demonstrated earlier, resonance features generated in the time-frequency domain from an echo backscattered by a target in free space,4–6 or buried underground,7 when a pulse from an impulse radar is incident on it, could be used to identify the target. A precondition in these time-frequency analyses is that the broad-band pulses emitted by the impulse radar do not exhibit distorting anomalies in their spectra. These anomalies could detrimentally affect the extracted resonance features in the time-frequency distribution of the pulses backscattered by any target.

In the implementation of E-pulse technique based radar target recognition system, only one E-pulse is constructed for each target and it is considered to be aspect-independent, which is grounded in the knowledge that E-pulse synthesis is based on the poles of targets and they are aspect-independent. But our recent study suggests that the poles obtained by pole extraction algorithm are aspect-dependent although theoretically poles of targets are aspect-independent. This is because the residues of the poles are aspect-dependent. The residues represent the contributions of the poles to the late-time transient scattered field and they may be so small at some aspect angles that the corresponding poles can not be extracted by pole extraction algorithm. Consequently, different poles may be extracted at different aspect angles. If only one E-pulse is constructed using the poles extracted at one aspect angle, it may not work for the transient scattered fields at other aspect angles. In order to solve this problem, we propose a multi-directional E-pulse technique which constructs multiple E-pulses for different angle regions of a target. Numerical results show that the proposed method is very robust and can dramatically improve the recognition rate of E-pulse technique.

The scattered field from a ramp driving function traces out the transverse cross sectional area profile of a metallic target as a function of time or depth of the target. This concept has been used in the past to generate 'low frequency' images of simple conducting shapes. This concept is currently being examined for dielectric scatterers such as land mines. It has been applied to theoretically computed scattered fields of dielectric spheres and to experimentally measured scattered fields from dielectric targets in free space. There are some advantages to applying this technique to identify land mines. First, the ramp response yields information about target size, orientation and geometrical shape. Second, the early time component of the scattered fields used in this imaging concept is relatively strong as contrasted to the late time response used for resonant target identification. Third, it is a low frequency concept since the ramp response frequency spectrum is proportional to the impulse response spectrum weighted by the inverse of the frequency squared. This is significant since the attenuation in most earth at a given depth increases as frequency increases for the frequency bandwidth used to detect land mines. However, for dielectric targets, the critical angle and internal scattering mechanisms can change the ramp response from their transverse cross sectional areas versus distance. Physical Optics solutions for the ramp response of dielectric spheres have been obtained and have been compared with their corresponding eigen function solutions. For a dielectric target, a way to improve the ramp response in the presence of the internal specular scattering mechanisms has been discussed. Also the ramp response signatures of some dielectric targets including a land mine kept in free space have been extracted from measured data for broadside and endfire excitations.

We have computed transient borehole electromagnetic (EM) responses of two-dimensional (2-D) models using a direct and explicit finite-difference algorithm. The program computes the secondary electric field which is defined as the difference between the total field and the primary (half-space) field. The time derivative of the vertical magnetic field in a borehole is computed by numerical differentiation of the total electric field. These models consist of a thin horizontal conductor with a finite width, embedded in a conductive half-space. Dual line sources energized by a step-function current lie on the surface of the half-space and simulate the long sides of a large rectangular loop.Numerical results substantiate several important features of the transient impulse response of such models. The peak response of the target is attenuated as the resistivity of the host decreases. A sign reversal in the secondary electric field occurs later in time as the resistivity of the host decreases. The peak response and the onset of late-time behavior are delayed in time as well. Secondary responses for models with different host resistivities (10-1000 Omega -m) are approximately the same at late time. If the target is less conductive, the effects of the host, i.e., the attenuation and time delay, are less. It is readily apparent that there exists a time window within which the target's response is at a maximum relative to the half-space response. At late time the shape of the borehole anomaly due to a thin conductive 2-D target appears to be independent of the conductivity of the host. The late-time secondary decay of the target is neither exponential nor power law, and a time constant computed from the slope of a log-linear decay curve at late time may be much larger than the actual value for the same target in free space.

The effect of linear polarization including E-polarization and H-polarization of the incident wave on the behaviour of NRCS of targets in free space has been analyzed numerically. The scattering problems of plane wave and beam wave incidences were considered. Target configuration together with beam width kW have major effects on the backscattering enhancement in the laser RCS (LRCS). Creeping waves, produced in the case of H-polarization, influence the LRCS obviously for limited ka and their impact diminishes gradually with ka. NRCS suffers from an obvious oscillated behavior in the range ka > kW and differs with polarization. NRCS decreases gradually with ka as a result of the lack in the surface current. NRCS approaches certain values with ka > kW irrespective of the linear wave polarization.

In a previous study, the radar cross-section (RCS) was shown to be influenced largely by the curvature of the illumination region of the target. Scattering data from smooth concave–convex contours reveals the obvious impact of specular reflections on the behavior of RCS. This effect needs to be investigated more especially with relatively complex surfaces. Here, we work on a numerical calculation of the RCS and analyze its characteristics with different target configurations including complexity and size. We postulate a concave illumination region and consider targets that are taking large sizes of about five wavelengths. In this communication, we assume wave propagation and scattering from targets in free space and random medium with consideration of horizontal incident wave polarization.

Accommodative state of young adults using reading spectacles

Synthetic aperture radar (SAR) target recognition under extended operating conditions (EOCs) is a challenging problem due to the complex application environment, especially for insufficient target variations and corrupted SAR images in the training samples. This paper proposes a new strategy to solve these problems for target recognition. The SAR images are firstly characterized by multi-scale components of monogenic signal. The generated monogenic features are decomposed to learn a class dictionary and a shared dictionary, which represent the possible intraclass variations information and the common information, respectively. Moreover, a sparse representation of the class dictionary and a dense representation of the shared dictionary are jointly employed to represent a query sample for classification. The validity of the proposed strategy is demonstrated with multiple comparative experiments on moving and stationary target acquisition and recognition (MSTAR) database.

In recent years, the micro-Doppler effect has been proposed as a new technique for signature analysis and extraction of radar targets. The ballistic missile is known as a typical radar target and has been paid many attentions for the complexities of its motions in current researches. The trajectory of a ballistic missile can be generally divided into three stages: boost phase, midcourse phase and terminal phase. The midcourse phase is the most important phase for radar target recognition and interception. In this stage, the warhead forms a typical micro-motion called micro-nutation which consists of three basic micro-motions: spinning, coning and wiggle. This paper addresses the issue of signature analysis of ballistic missile warhead in terahertz band via discussing the micro-Doppler effect. We establish a simplified model (cone-shaped) for the missile warhead followed by the micro-motion models including of spinning, coning and wiggle. Based on the basic formulas of these typical micro-motions, we first derive the theoretical formula of micro-nutation which is the main micro-motion of the missile warhead. Then, we calculate the micro-Doppler frequency in both X band and terahertz band via these micro-Doppler formulas. The simulations are given to show the superiority of our proposed method for the recognition and detection of radar micro targets in terahertz band.

Images from different aspect views for one target, known as multiple views, are widely applied to improve synthetic aperture radar (SAR) target recognition. However, most of existing multi-view methods have strict constraint on the angle interval among multiple views. In this paper, a new multiview fusion method free from interval limitation using expectation maximization (EM) is explored for SAR image classification. Firstly, we apply convolutional neural network (CNN) to extract features effectively owning to its powerful ability of feature learning and then obtain the classification probability. Secondly, Multi-view Label Set (MLS) is automatically constructed from multiple views according to the probability and finally we use EM algorithm to classify SAR images intelligently. It is worth noting that the proposed method can be used flexibly according to the number of perspectives obtained and without angle interval constraint among multiple views. Experiments demonstrate that the proposed method has better recognition performance than some state-of-the-art methods on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset.

A synthetic aperture radar (SAR) target recognition method was proposed based on multivariate empirical mode decomposition (MEMD). MEMD was the general extension of traditional EMD, which could avoid the mode mixing problems. MEMD was employed to process SAR images to obtain the multi-layer intrinsic mode functions (IMF), which could better reflect the time-frequency properties of the targets. Different layers of IMFs could effectively complement each other while sharing some inner correlations because they are generated from the same target. In the classification phase, the joint sparse representation was employed to represent the IMFs. The joint sparse representation could solve several related sparse representation tasks based on the idea of multi-task learning. It could produce more precise estimations than the solutions of single tasks. According to the sparse coefficient vectors corresponding to different IMFs, the reconstruction errors of different classes for the representation of the test sample can be calculated. Afterwards, the target label of the test sample can be determined. Experiments were conducted on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset, by comparison with existing methods under the standard operating condition, depression angle variance, noise corruption, and target occlusion, the results confirm the validity of the proposed method.

In order to improve synthetic aperture radar (SAR) target recognition performance, this paper proposes a novel method using multi-level deep features. The multi-level deep features are learned by the convolutional neural network (CNN), which are capable of describing the target characteristics from different aspects. In order to make full use of the discrimination contained in the multi-level deep features, the joint sparse representation (JSR) is used as the basic classifier, which performs the multi-task learning to jointly classify the multi-level deep features. It could not only represent each feature properly but also consider the correlations between different levels of features. Based on the solutions from JSR, the target label is classified as the training class with the minimum reconstruction error. By fully exploiting the discriminative information contained in the multi-level deep features, the proposed method could effectively enhance SAR target recognition performance. The moving and stationary target acquisition and recognition (MSTAR) dataset is employed in the experiments. The results show that the proposed method could achieve a significantly high recognition rate of 99.38% for classifying 10 classes of targets under the standard operating condition (SOC), which is higher than those from some reference methods drawn from current literatures. Under different types of extended operating conditions (EOCs), the overall performance of the proposed method keeps superior over the reference methods. In addition, the outlier rejection capability of the proposed method is also better than the compared methods. All these experimental results validate the high effectiveness of the proposed method.