Detection of Landmines from Acoustic Images Based on Cepstral Coefficients

This paper introduces a cepstral approach for the automatic detection of landmines from acoustic images. This approach is based on treating the problem of landmine detection as a pattern recognition problem. Cepstral features are extracted from a group of landmine images which are transformed first to 1-D signals by lexicographic ordering. Mel frequency cepstral coefficients (MFCCs) and polynomial shaping coefficients are extracted from these 1-D signals to form a database of features, which can be used to train a neural network with the landmine features. The landmine detection can be performed by extracting features from any new image with the same method used in the training phase. These features are tested with the neural network to decide whether a landmine exists or not. The different domains are tested and compared for efficient feature extraction from the lexicographically ordered 1-D signals. Experimental results show the success of the proposed cepstral approach for landmine detection at low as well as high signal to noise ratios. Results also show that the discrete cosine transform is the most appropriate domain for feature extraction.     

Clutter Reduction and Detection of Minelike Objects in Ground Penetrating Radar Data Using Wavelets

Ground Penetrating Radar (GPR) signatures of shallowly buried landmines are normally obscured by a strong background signal comprised of the reflections from the ground surface and the antenna crosstalk. Based on the notion that buried landmine produces an anomaly in the soil dielectric an automated procedure has been developed which detects soil dieletric anomalies of the size comparable to the size of landmine in GPR data and enhances the signatures of such anomalies. A local background estimate is computed and a soil dielectric anomaly is detected at the spatial position where a change from the estimated background signal occurs. A translation invariant wavelet packet decomposition is applied for detection. The computation takes place in a running window which allows for the algorithm to adapt to the variations in ground conditions and antenna height. The technique was tested using a number of minelike targets buried in several different soil environments and the testing results are presented.     

Performance Comparison of Automated Induction-Based Algorithms for Landmine Detection in a Blind Field Test

Wideband Electromagnetic Induction (EMI) data provides an opportunity to apply robust statistical signal processing techniques to potentially mitigate false alarm rates in real-time landmine detection. This paper explores the application of matched subspace detectors (MSDs) and Support Vector Machines (SVMs) to this problem. A library of landmine responses is generated from a set of calibration data and a bank of matched subspace detectors, each designed to detect a specific mine type, is developed. Support vector machines are also considered for target/clutter discrimination. These are developed based on landmine signatures, decay rate estimates, and the outputs of matched subspace filter banks. Synthetic data sets are generated and matched subspace detectors and support vector machines are trained using this synthetic data. Receiver Operating Characteristics (ROCs) for matched subspace detectors and support vector machines are presented for both experimental and simulated data sets. The results indicate that substantial reductions in false alarm rates can be achieved using these techniques, but that simulated data sets may not provide accurate predictors of performance.     

Analysis of Linear Prediction for Soil Characterization in GPR Data for Countermine Applications

Ground-penetrating radar (GPR) is a versatile technology for subsurface sensing, and has shown promise in countermine applications when a target detection algorithm is employed. Because the soil environment is naturally heterogeneous and nonstationary, many detection algorithms have taken the form of adaptive filters operating on the real-aperture radar data. In particular, linear prediction techniques have received much attention for their ability to screen for anomalous signals that differ from the background. In this work, we demonstrate that linear prediction may provide a low-dimensional feature set that is indicative of various soil properties. Experiments were performed with simulated and field-collected GPR data, and results provide greater understanding of how linear predictors might be useful in landmine detection over varying terrain.     

Dual-Frequency Microwave-Enhanced Infrared Thermography

In this paper, we propose to extend the Microwave Enhanced Infrared Thermography (MEIT) method for landmine detection using two excitation microwave frequencies. We present the results of a detailed modeling effort. The extended method has the potential for detecting buried objects under rough surfaces at depths greater than with the original single frequency concept. Our result shows that the two-frequency technique minimizes the clutter introduced by an irregular surface, and can lead to better detection of buried objects. A 2-D computational model of this method has been developed to simulate real-world landmine detection. Receiver Operating Characteristic (ROC) curves were generated to evaluate the performance of this new method. Finally we demonstrate improvements achieved in detection depth under rough surfaces, compared to the single-frequency MEIT technology. The theoretical results obtained so far are presently being used to plan laboratory and field experiments for validation of our assumptions.     

Data Processing and Imaging in GPR System Dedicated for Landmine Detection

Improvements of GPR technology can be attained by making adjustments specific for the application of landmine detection on three levels: system design, data acquisition and data processing. In this paper we describe data processing algorithms specially developed for a novel video impulse ultra-wide band front end. With this front end, three-dimensional measurements (C-scans) have been carried out over a controlled test site, using a non-metallic scanner. The test site contained surface-laid and shallow buried landmines, both antitank and antipersonnel, made of plastic, wood, and metal. Because of practical limitations, the data have been acquired on an irregular grid. We have designed data preprocessing and imaging algorithms such that they take into account the specific antenna geometry and its elevation above the ground as well as the irregularity of the data acquisition grid. We show that by tuning the data pre-processing and imaging to the newly designed radar front end and to the particular data acquisition strategy, we obtain clear subsurface images. The resulting images show the ability of the GPR system to detect and visualize small surface laid and shallow buried targets.     

An investigation into diesel engine air-borne acoustics using continuous wavelet transform

The emitted acoustic signals from diesel engines carry useful indicators about their operating conditions and health status. Unfortunately, those signals are very complex, contain numerous numbers of sources and corrupted by subnational amount of noise. This makes it difficult to extract those condition indicators via the use of conventional time and frequency domain analysis techniques. This paper studies the characteristics of diesel engine air-borne acoustic signals using time-frequency domain techniques. One analysis technique is investigated; continuous wavelet transform (CWT). First, some of the mathematical background of the CWT is reviewed. Second, the detection capabilities of this technique are evaluated using air-borne acoustic signals collected from diesel engine in acoustically untreated laboratory. Consequently, some engine operating conditions and faults are investigated using the CWT techniques. The achieved results prove the technique’s sensitivity to engine speed and load variations. More important the CWT shows excellent capabilities in detecting engine’s injection process and lubrication related faults at early stages.     

A comparison study of support vector machines and hidden Markov models in machinery condition monitoring

Condition classification is an important step in machinery fault detection, which is a problem of pattern recognition. Currently, there are a lot of techniques in this area and the purpose of this paper is to investigate two popular recognition techniques, namely hidden Markov model and support vector machine. At the beginning, we briefly introduced the procedure of feature extraction and the theoretical background of this paper. The comparison experiment was conducted for gearbox fault detection and the analysis results from this work showed that support vector machine has better classification performance in this area.     

Enhanced Auditory Displays for Discriminating Landmines from Clutter Using Electromagnetic Induction Sensors

Landmine detection is primarily performed using electromagnetic induction (EMI) sensors. These sensors detect the presence of metal and convey the information to the sensor operators via an audio signal. Reduction of false alarms from objects that contain metal but are not landmines, i.e. discrimination, is a challenging problem. Recent work on automated algorithms has shown promise towards reducing false alarm rates of EMI sensors. In this study, the audio signal was modified to encode the presence of metal as well as information regarding mine/non-mine belief in order to determine whether the additional information enabled operators to better discriminate mines from clutter. Using data collected from real landmines, we experimentally investigated which perceptual dimensions most effectively convey different aspects of the information contained in the sensor response to a listener. Results indicated that the presence of metal (detection) could be coded in the fundamental frequency of the audio signal, and that mine/non-mine belief (discrimination), determined using an automated algorithm, could be coded in a separate audio dimension. Operators performed better with this audio coding scheme than one where only metal content information was presented via the fundamental frequency of the audio signal.     

An observer based approach for achieving fault diagnosis and fault tolerant control of systems modeled as hybrid Petri nets

In this paper, we propose an approach for achieving detection and identification of faults, and provide fault tolerant control for systems that are modeled using timed hybrid Petri nets. For this purpose, an observer based technique is adopted which is useful in detection of faults, such as sensor faults, actuator faults, signal conditioning faults, etc. The concepts of estimation, reachability and diagnosability have been considered for analyzing faulty behaviors, and based on the detected faults, different schemes are proposed for achieving fault tolerant control using optimization techniques. These concepts are applied to a typical three tank system and numerical results are obtained.     

以内容为主的影片检索摘要技术及其在网络多媒体出版上之应用

目前根据影片结构及MPEG影片压缩技术所发展出来的分镜侦测技术目前已有相当多种可供利用,举例来说在未压缩域有pixel-wise comparison,likelihood ratio及histogram comparison;而在压缩域则有MPEG离散余弦系数及直流值画面等方法.利用分镜侦测技术所发展出来的数字影片内容分析检索技术可应用在许多的领域,可节省在网络及一般影片出版流程上人力所需耗时处理的动作.为此针对几项主要的分镜侦测方法的原理及其优缺点进行说明,并以网络影片数据库及自动化的内容查询系统两项例子来说明其在网络多媒体出版上的应用.     

DAMAGE ASSESSMENT USING MODE SHAPE SENSITIVITIES

In this paper, a method for the localisation and assessment of damage is presented. The method is based on the use of mode shape sensitivities to changes in mass or stiffness in the test structure. With these sensitivities, differences in the dynamical behaviour of the structure in its undamaged and damaged conditions can be translated into damage information (location and amount of changes in mass or stiffness). Since the sensitivities are calculated on the basis of the experimentally determined mode shapes, there is no need for a prior finite element model of the test structure. The applicability of this technique is discussed on the basis of damage detection experiments performed on a beam-like structure (laboratory conditions) as well as data from the experiments on the I-40 highway bridge in Albuquerque, New Mexico. A comparison is made with a variety of existing damage detection techniques which do not require a finite element model.     

声波探雷及其信号处理方法研究

为了研究声波探雷机理,设计一个基于声-地震耦合的实验系统,并用曲线拟合的方法研究复杂的测试信号变化规律。首先,简述声-地震耦合基本原理,分析土壤-地雷系统的共振作用;然后,分别用喇叭和地震检波器串激发和接收地表振动;最后,采用平均值算法进行地表振动速度曲线拟合,并解释实验现象。结果显示有地雷的地表振动速度与无雷的比值在很宽的频带内都大于1,其峰值达24。表明地雷能引起地表振动的明显变化,所提出的实验系统及曲线拟合方法可用于声波探雷信号的进一步研究。     

甲烷检测技术的研究现状

甲烷的安全防爆是矿山开采中的一个重要问题,甲烷浓度实时准确的检测是问题的关键。本文重点阐述两类基于不同原理的甲烷检测技术:光纤吸收法、光干涉法。并对光纤吸收法中的差分检测技术、谐波检测技术进行比较,对各种方法检测甲烷的检测精度进行分析。     

Commercial Systems for the Direct Detection of Explosives for Explosive Ordnance Disposal Tasks

There are a number of Explosive Ordnance Disposal (EOD) tasks in which it is necessary to discriminate and identify UneXploded Ordnance (UXO) already detected by other means. What we seek is the capacity to characterize in a non-destructive way the munition's content, usually either explosive or inert (e.g., practice munition), if possible using a fairly mobile system and without direct contact. A number of existing technologies for the direct detection of explosives, applicable in EOD scenarios, are therefore characterized, and corresponding commercially available systems or advanced prototypes identified whenever possible. Some of the techniques are also useful for the detection of buried UXO or landmines, to which we refer whenever appropriate. We first concentrate on bulk explosive detection, in particular on neutron based systems exploiting gamma spectroscopy, which have the potential of detecting the explosive's nitrogen content and/or its other constituents (carbon, oxygen and hydrogen). Candidate systems exist, although most of them have as primary aim the discrimination of chemical munition. We then describe the application of trace explosive detection techniques, which seems to be less mature also due to the large number of parameters influencing the variables of interest (explosive vapor and particle concentration). Sampling is in this case of primary importance.     

制造过程产品检测信息建模技术研究

阐明了产品检测信息建模的目的和意义,介绍了产品检测信息建模技术研究的现状,分析了产品检测信息建模技术的一般性要求和特殊性要求。由此出发,应用并行工程原理,采用人工智能与人机一体化和面向对象技术,建立了相应的检测信息获取模型和表达模型。     

EXPLORING ONCE-PER-REVOLUTION AUDIO SIGNAL VARIANCE AS A CHATTER INDICATOR

The purpose of this study is an evaluation of the statistical variance in the once-per-revolution sampled audio signal during milling as a chatter indicator. It is shown that, due to the synchronous and asynchronous nature of stable and unstable cuts, respectively, once-per-revolution sampling leads to a tight distribution of values for stable cuts, with a corresponding low variance, and a wider sample distribution for unstable cuts, with an associated high variance. A comparison of stability maps developed using: 1) analytic techniques, and 2) the variance from once-per-revolution sampled time-domain simulations is provided and good agreement is shown. Experimental agreement between the well-known Fast Fourier Transform (FFT) chatter detection method, that analyzes the content of the FFT spectrum for chatter frequencies, and the new variance-based technique is also demonstrated.     

EXPLORING ONCE-PER-REVOLUTION AUDIO SIGNAL VARIANCE AS A CHATTER INDICATOR

The purpose of this study is an evaluation of the statistical variance in the once-per-revolution sampled audio signal during milling as a chatter indicator. It is shown that, due to the synchronous and asynchronous nature of stable and unstable cuts, respectively, once-per-revolution sampling leads to a tight distribution of values for stable cuts, with a corresponding low variance, and a wider sample distribution for unstable cuts, with an associated high variance. A comparison of stability maps developed using: 1) analytic techniques, and 2) the variance from once-per-revolution sampled time-domain simulations is provided and good agreement is shown. Experimental agreement between the well-known Fast Fourier Transform (FFT) chatter detection method, that analyzes the content of the FFT spectrum for chatter frequencies, and the new variance-based technique is also demonstrated.     

Characterization of defects in semiconductors by combined application of SEM(EBIC) and SDLTS*

Besides the characterization of the geometrical structure of defects in semiconductors by TEM the estimation of their electrical activity is of importance. SEM(EBIC) and SDLTS (scanning deep level transient spectroscopy) are especially suitable for this purpose; they allow the inspection of electronic properties with a spatial resolution in the micron-range. On the one hand, SEM(EBIC) yields information on the recombination efficiency of defects in the crystal volume adjacent to a pn junction or a Schottky barrier; on the other hand, SDLTS enables the detection to be carried out of the distribution and the energetic levels of deep level defects lying in the space charge region. Accordingly, the combined application of these techniques is very promising for investigating physical processes implying an inhomogeneous incorporation of deep level defects in semiconductor crystals. In comparison to the widely used SEM(EBIC) technique SDLTS has only rarely been applied, a fact that is due to the high detection sensitivity necessary for measuring capacity transients. The application of a highly sensitive (10^—6 pF) micro-computer-controlled SDLTS system in combination with a conventional EBIC system allows a reliable inspection of semiconductor materials and devices, based on A₃B₅ compounds and on silicon. A typical application of the above technique is the investigation of the impurity distribution around extended crystal defects, like dislocations and precipitates, to study their gettering activity.     

Transient-spatial pattern mining of eddy current pulsed thermography using wavelet transform

Eddy current pulsed thermography（ECPT） is an emerging Non-destructive testing and evaluation（NDT E） technique, which uses hybrid eddy current and thermography NDT E techniques that enhances the detectability from their compensation. Currently, this technique is limited by the manual selection of proper contrast frames and the issue of improving the efficiency of defect detection of complex structure samples remains a challenge. In order to select a specific frame from transient thermal image sequences to maximize the contrast of thermal variation and defect pattern from complex structure samples, an energy driven approach to compute the coefficient energy of wavelet transform is proposed which has the potential of automatically selecting both optimal transient frame and spatial scale for defect detection using ECPT. According to analysis of the variation of different frequency component and the comparison study of the detection performance of different scale and wavelets, the frame at the end of heating phase is automatically selected as an optimal transient frame for defect detection. In addition, the detection capabilities of the complex structure samples can be enhanced through proper spatial scale and wavelet selection. The proposed method has successfully been applied to low speed impact damage detection of carbon fibre reinforced polymer（CFRP） composite as well as providing the guidance to improve the detectability of ECPT technique.