Target Classification Method Research Articles

Target Detection and Classification Using Seismic and PIR Sensors

Unattended ground sensors (UGS) are widely used to monitor human activities, such as pedestrian motion and detection of intruders in a secure region. Efficacy of UGS systems is often limited by high false alarm rates, possibly due to inadequacies of the underlying algorithms and limitations of onboard computation. In this regard, this paper presents a wavelet-based method for target detection and classification. The proposed method has been validated on data sets of seismic and passive infrared sensors for target detection and classification, as well as for payload and movement type identification of the targets. The proposed method has the advantages of fast execution time and low memory requirements and is potentially well-suited for real-time implementation with onboard UGS systems.

Radar target classification method with high accuracy and decision speed performance using MUSIC spectrum vectors and PCA projection

This paper introduces the performance of an electromagnetic target recognition method in resonance scattering region, which includes pseudo spectrum Multiple Signal Classification (MUSIC) algorithm and principal component analysis (PCA) technique. The aim of this method is to classify an “unknown” target as one of the “known” targets in an aspect‐independent manner. The suggested method initially collects the late‐time portion of noise‐free time‐scattered signals obtained from different reference aspect angles of known targets. Afterward, these signals are used to obtain MUSIC spectrums in real frequency domain having super‐resolution ability and noise resistant feature. In the final step, PCA technique is applied to these spectrums in order to reduce dimensionality and obtain only one feature vector per known target. In the decision stage, noise‐free or noisy scattered signal of an unknown (test) target from an unknown aspect angle is initially obtained. Subsequently, MUSIC algorithm is processed for this test signal and resulting test vector is compared with feature vectors of known targets one by one. Finally, the highest correlation gives the type of test target. The method is applied to wire models of airplane targets, and it is shown that it can tolerate considerable noise levels although it has a few different reference aspect angles. Besides, the runtime of the method for a test target is sufficiently low, which makes the method suitable for real‐time applications.

표적의 기하학적 변환에 강인한 SIFT 기반의 표적 분류 알고리즘 설계

This paper proposes a method for classifying targets robust to geometric transformations of targets such as rotation, scale change, translation, and pose change. Targets which have rotation, scale change, and shift is firstly classified based on CM(Confidence Map) which is generated by similarity, scale ratio, and range of orientation for SIFT(Scale-Invariant Feature Transform) feature vectors. On the other hand, DB(DataBase) which is acquired in various angles is used to deal with pose variation of targets. Range of the angle is determined by comparing and analyzing the execution time and performance for sampling intervals. We experiment on various images which is geometrically changed to evaluate performance of proposed target classification method. Experimental results show that the proposed algorithm has a good classification performance.

An Electromagnetic Target Classification Method for the Target Sets with Alien Target: Application to Small-scale Aircraft Targets

The electromagnetic target classiflcation is a challenging problem since the scat- tered fleld from a target is highly dependent on operating frequency, polarization and aspect angle. In order to minimize adverse efiects of these dependencies an intelligent classifler con- taining some distinguishable target features is needed. In addition, in order to be suitable for real target applications, the properties of operating with moderate frequency bandwidth and discriminating an alien target from a target set containing friend targets are important. In this study, an electromagnetic target classiflcation method for isolated targets using noisy data in the classifler design to obtain high accuracy performance in a wider SNR range and having the ability of discrimination of an alien target without any priori information is introduced. The proposed method is mainly based on a late-time resonance region target classiflcation technique, which was reported recently to use the multiple signal classiflcation (MUSIC) algorithm and natural-resonance mechanism modeled by singularity expansion method (SEM) for target fea- ture extraction, and modifled for target sets containing alien target(s). The proposed classifler design method is demonstrated and tested for a target set of flve friend and one alien small-scale aircraft targets. According to the test result, the proposed method gives high accuracy rates for this target set.

Radar target classification method with reduced aspect dependency and improved noise performance using multiple signal classification algorithm

This study introduces a novel aspect and polarisation invariant radar target classification method based on the use of multiple signal classification (MUSIC) algorithm for feature extraction. In the suggested method, for each candidate target at each designated reference aspect, feature matrices called ‘MUSIC spectrum matrices (MSMs)’ are constructed using the target's scattered data at different late-time intervals. An individual MSM corresponds to a map of a target's natural resonance-related power distribution over the complex frequency plane under the chosen aspect angle/late-time interval conditions. The collection of these feature matrices is used first to determine the best late-time interval for optimal feature extraction. Then, the MSM of a target, which are computed over the optimal time interval at all reference aspects, are superposed to obtain the ‘fused MUSIC spectrum matrix (FMSM)’. The FMSM of a target is its main classifier feature in the proposed method as the aspect dependency of an FMSM is highly reduced because of its multi-aspect construction process. The suggested method is demonstrated for both simple and complex target geometries such as conducting spheres, dielectric spheres and small-scale aircraft targets with high accuracy rates even for low SNR values using feature fusion at only a few different reference aspects.

Maximum Position Alignment Method for Noisy High-Resolution Radar Target Classification

In this paper, the alignment of noisy high-resolution radar signals using the maximum position method is studied. The relationship between the shift estimation and the signal-to-noise ratio is considered. As a result, two analytical expressions are obtained that approximate the root-mean-square error of the difference in the shift estimation with and without noise. These two expressions allow us to improve the understanding of the sensitivity to noise of the Maximum Position alignment method.

Modeling and analysis of the backscatter of dolphin echolocation clicks from elastic targets

The clicks from an Atlantic bottlenose dolphin were obtained in a controlled experiment at the Marine Mammal Research Program (Kaneohe Bay, HI). A previously described phantom echo experiment using solid elastic spherical targets provides the foundation for an investigation of the specific scattering components used in a dolphin’s target identification and classification methods. A model of the scattering interaction of incident dolphin echolocation clicks and elastic targets is outlined. The frequency spectra from the echolocation clicks are combined with the resonance-scattering representations of the scattered pressure fields to yield filter-type integrals that can be viewed as filtering the form function of the scatterer through the spectral window of the incident echolocation clicks. This analysis is carried out in the time and frequency domains and also represented in time-frequency distributions (Wigner-Ville, RID). An advantage of this analysis is that it can be carried out for any arbitrary echolocation click incident on targets of any general shape and composition. The results are compared and discussed with respect to experimentally obtained backscattered signals.

Discriminating real objects in radar imaging by exploiting the squared modulus of the continuous wavelet transform

New technique based on continuous wavelet transform (CWT) for classifying objects in synthetic aperture radar (SAR) imaging is presented. The CWT allows to analyse two-dimensional SAR images to highlight the frequency and angular behaviour of the scatterers ref. 10, 11. This technique allows to build a SAR hyperimage, that is, a four-dimensional data cube which represents for each spatial location (x, y) of the scatterer in the image, its frequency and angular energy behaviour. When analysing different targets, objects or areas in SAR images, it has been recently observed that some scatterers belonging to a same class of objects could have similar frequency and angular energy responses. The previous observations have motivated the determination to exploit these energy responses to discriminate these objects. This discrimination is performed by frequency and angular correlations between the response of a particular scatterer (measured) and those of all the scatterers in the SAR image. Some examples of discrimination from real SAR data are presented and show an interest of the method for target classification and recognition for SAR imaging.

Undersea Target Classification Using Canonical Correlation Analysis

Canonical correlation analysis is employed as a multiaspect feature extraction method for underwater target classification. The method exploits linear dependence or coherence between two consecutive sonar returns, at different aspect angles. This is accomplished by extracting the dominant canonical correlations between the two sonar returns and using them as features for classifying mine-like objects from nonmine-like objects. The experimental results on a wideband acoustic backscattered data set, which contains sonar returns from several mine-like and nonmine-like objects in two different environmental conditions, show the promise of canonical correlation features for mine-like versus nonmine-like discrimination. The results also reveal that in a fixed bottom condition, canonical correlation features are relatively invariant to changes in aspect angle.

A Novel Application of a Microaccelerometer for Target Classification

This paper presents a novel method of target classification by means of a microaccelerometer. Its principle is that the seismic signals from moving vehicle targets are detected by a microaccelerometer, and targets are automatically recognized by the advanced signal processing method. The detection system based on the microaccelerometer is small in size, light in weight, has low power consumption and low cost, and can work under severe circumstances for many different applications, such as battlefield surveillance, traffic monitoring, etc. In order to extract features of seismic signals stimulated by different vehicle targets and to recognize targets, seismic properties of typical vehicle targets are researched in this paper. A technique of artificial neural networks (ANNs) is applied to the recognition of seismic signals for vehicle targets. An improved back propagation (BP) algorithm and ANN architecture have been presented to improve learning speed and avoid local minimum points in error curve. The improved BP algorithm has been used for classification and recognition of seismic signals of vehicle targets in the outdoor environment. Through experiments, it can be proven that target seismic properties acquired are correct, ANN is effective to solve the problem of classification and recognition of moving vehicle targets, and the microaccelerometer can be used in vehicle target recognition.

Multi-platform sonar concepts for buried target detection and classification

Buried target classification is of paramount importance in mine countermeasures (MCM) applications. The Generic Oceanographic Array Technology Sonar (GOATS) project approaches this fundamental problem by exploiting the assets of autonomous underwater vehicles (AUVs). Recent developments in unmanned vehicle technology have opened the door for new sonar design concepts that may be applied to create alternative target detection and classification methods in complex environments. Adaptive vehicle behaviors provide the potential for the sonar system to adjust its geometry for optimal performance. The vehicles can also access very shallow water environments, which allows for more complete multi-aspect views of targets of interest than are possible with remote towed array systems. In this paper, it is demonstrated that the multi-platform system can provide improved detection and classification capability, while at the same time reducing the computational requirements. Data from prior GOATS experiments (1998 and 2000) are used to illustrate the effectiveness of the multi-platform sonar concepts, and real-time simulations are used to show the planned implementation of these techniques on-board the AUVs for GOATS’02.

Joint target tracking and classification using radar and ESM sensors

Bayesian target classification methods using radar and electronic support measure (ESM) data are considered. A joint treatment of target tracking and target classification problems is introduced. First, a method for target classification using radar data and class-dependent kinematic models is presented. Second, a target classification method using ESM data is presented. Then, a Bayesian radar and ESM data fusion algorithm, referred to as direct identity fusion (DIF), for target classification is presented. This algorithm exploits the dependence of target state on target class via the use of class dependent kinematic models but fails to exploit the dependence of target class on target state. We then introduce a method, referred to as joint tracking and classification (JTC), for treating target tracking and classification problems jointly, by exploiting the dependence of target class on target state via flight-envelope-dependent classes and the dependence of target state on target class via class dependent kinematic models. A two-dimensional example demonstrates the relative merits of these methods. It is shown that, while the incorporation of the two-way dependence between target state and class (i.e., JTC) promises some benefits over the method that incorporates only a one-way dependence (i.e., DIF), there are severe filter implementation difficulties for the former. The results also demonstrate that the fusion of information from radar and ESM sensors via the DIF approach results in improvements over classification methods based on either of the individual sensors.

Multiple neural networks-integrated underwater target classification based on fuzzy theory

In the area of underwater target recognition, in order to process various target feature data extracted from underwater radiating noise signals, a target classification method based on the integration of multiple fuzzy clustering neural networks and fuzzy comprehensive analysis is put forward in this paper. An adaptive fuzzy clustering neural network classifier is constructed and a relative studying algorithm is described for the first class classification, which is applied separately to different kinds of feature data from LOFAR spectrum, bispectrum, and wavelet transformation. Based on fuzzy comprehensive analysis of the outputs and probabilities of the above neural network classifiers and the category and distribution of test samples, the second class fuzzy classifying rules are designed to preprocess and assess the categories of ships. More than 4000 passive sonar signal samples are collected and many tests are done, and the results show the feasibility, adaptability, and effectiveness of the present method.

A neural network harmonic family classifier

A method of automatic target discrimination and classification for passive sonar applications is obtained using an unsupervised, self-organizing, Kohonen neural network. Utilizing detected, smoothed spectral data, the kth-order frequency differences are histogrammed into constant percentage resolution spectral bins prior to presentation to the network. This nonlinear transformation results in good neural network classification when the input data is topologically close to the training set. Simulations of the algorithm show highly robust performance to missing spectral detections, spurious spectral detections and Doppler shifts.

A Neural Network Method for High Range Resolution Target Classification

A Neural Network Method for High Range Resolution Target Classification