Explore the vast range of titles available.

Deep learning-based sustainable subsurface for anomaly detection in different materials is an objective to improve the reliability of thermographic inspection. This article aims to describe a method that uses Barker-coded thermal wave imaging to identify subsurface anomalies in materials. The novelty of the proposed methodology is to detect smaller defects at a higher depth even on a fully corroded sample of mild steel. Experiments were carried out with different kinds of samples like mild steel and glass fiber reinforced plastic (GFRP). Various commonly used modern post-processing techniques are applied alongside the proposed techniques for detecting subsurface anomalies. Subsurface anomalies visualized using the proposed deep learning method give better visualization and results when compared to that of other approaches. In addition to it, region-based active contour segmentation-based detection is also proposed for the GFRP sample. This study results in a high signal-to-noise ratio (SNR) of value 108 dB; the least error in defect size is nearly 0.01% using full width at half maximum (FWHM), and the aspect ratio is nearly 1 for the proposed convolutional neural network (CNN)-based processing approach.

PurposeIn order to include the non-negligible lag relaxation time feature that is characteristic of heat transfer in biological bodies, the classical Fourier's law of heat conduction has to be generalized as the Maxwell–Cattaneo law resulting in the thermal-wave model of bio-heat transfer. The purpose of the paper is to retrieve the unknown time-dependent blood perfusion coefficient in such a thermal-wave model of bio-heat transfer from (non-intrusive) measurements of the temperature on an accessible sub-portion of the boundary that may be taken with an infrared scanner.Design/methodology/approachThe nonlinear and ill-posed problem is reformulated as a nonlinear minimization problem of a Tikhonov regularization functional subject to lower and upper simple bounds on the unknown coefficient. For the numerical discretization, an unconditionally stable direct solver based on the Crank–Nicolson finite-difference scheme is developed. The Tikhonov regularization functional is minimized iteratively by the built-in routine lsqnonlin from the MATLAB optimization toolbox. Numerical results for a benchmark test example are presented and thoroughly discussed, shedding light on the performance and effectiveness of the proposed methodology.FindingsThe inverse problem of obtaining the time-dependent blood perfusion coefficient and the temperature in the thermal-wave model of bio-heat transfer from extra boundary temperature measurement has been solved. In particular, the uniqueness of the solution to this inverse problem has been established. Furthermore, our proposed computational method demonstrated successful attainment of the perfusion coefficient and temperature, even when dealing with noisy data.Originality/valueThe originalities of the present paper are to account for such a more representative thermal-wave model of heat transfer in biological bodies and to investigate the possibility of determining its time-dependent blood perfusion coefficient from non-intrusive boundary temperature measurements.

The active thermal non-destructive testing and evaluation technique plays a vital role in health monitoring of various solid materials. Present manuscript demonstrates the applicability of pulse compression favorable Digitized version of linear Frequency Modulated Thermal Wave Imaging (DFMTWI) approach to identify flaws having different geometrical shapes in a Glass Fibre Reinforced Polymer (GFRP) sample. A novel Thermal Image Correlation (TIC) data-processing approach is proposed to obtain the isothermal patterns from the reconstructed pulse compressed data through matched filter scheme to identify sub-surface anomalies. The detection capabilities of the presented approach are compared on various adopted data processing approaches.

PurposeWhen modeling heat propagation in biological bodies, a non-negligible relaxation time (typically between 15-30 s) is required for the thermal waves to accumulate and transfer, i.e. thermal waves propagate at a finite velocity. To accommodate for this feature that is characteristic to heat transfer in biological bodies, the classical Fourier's law has to be modified resulting in the thermal-wave model of bio-heat transfer. The purpose of the paper is to retrieve the space-dependent blood perfusion coefficient in such a thermal-wave model of bio-heat transfer from final time temperature measurements.Design/methodology/approachThe non-linear and ill-posed blood perfusion coefficient identification problem is reformulated as a non-linear minimization problem of a Tikhonov regularization functional subject to lower and upper simple bounds on the unknown coefficient. For the numerical discretization, an unconditionally stable direct solver based on the Crank–Nicolson finite difference scheme is developed. The Tikhonov regularization functional is minimized iteratively by the built-in routine lsqnonlin from the MATLAB optimization toolbox. Both exact and numerically simulated noisy input data are inverted.FindingsThe reconstruction of the unknown blood perfusion coefficient for three benchmark numerical examples is illustrated and discussed to verify the proposed numerical procedure. Moreover, the proposed algorithm is tested on a physical example which consists of identifying the blood perfusion rate of a biological tissue subjected to an external source of laser irradiation. The numerical results demonstrate that accurate and stable solutions are obtained.Originality/valueAlthough previous studies estimated the important thermo-physical blood perfusion coefficient, they neglected the wave-like nature of heat conduction present in biological tissues that are captured by the more accurate thermal-wave model of bio-heat transfer. The originalities of the present paper are to account for such a more accurate thermal-wave bio-heat model and to investigate the possibility of determining its space-dependent blood perfusion coefficient from temperature measurements at the final time.

Non-Destructive Testing and Evaluation (NDT&E) is being developed across various segments of the industry for detecting the presence of defects occurring either during the manufacturing or its in-service stage such as cracks, voids, delamination, etc. in a wide variety of materials. Among various NDT&E methodologies, InfraRed Thermography (IRT) gained importance due to its remote, whole-field, safe, and quantitative assessment of industrial and biomaterials. These merits make the IRT a promising approach for inspecting and evaluating various composite structures widely used in aerospace and defense applications. Among the recently introduced IRT techniques, Frequency Modulated Thermal Wave Imaging (FMTWI) ensures the feasibility of implementing moderate peak power heat sources in single experimentation compared to conventional pulse-based and sinusoidally modulated lock-in thermography. The present work enhances the scope of the Principal Component Analysis (PCA)-based IRT named Principal Component Thermography (PCT) by pioneering the application of temporally and spatially reconstructed FMTWI dataset for the first time. This paper explores the PCT-based data processing algorithm to test and evaluate artificially simulated blind hole defects in a Glass Fibre Reinforced Polymer (GFRP) material. The results of PCT obtained for the FMTWI technique highlight the merits of the data-reduced feature map known as Empirical Orthogonal Thermogram (EOT) along with its defect detection capabilities by considering the optimal Principal Component (PC) to reduce the effect of uneven heating on the sample.

Deeper defect detection and depth resolution capabilities of quadratic frequency-modulated optical stimulus became a viable approach for material inspection in active infrared non-destructive testing modality. But the limitations of complex and non-linear analytical models associated with processing techniques propel towards automated defect assessment techniques in infrared thermography. This paper introduces a deep neural network-based automatic defect detection and depth visualization technique in quadratic frequency modulated thermal wave imaging. The neural network classifier uses the modified loss function of a one-class support vector machine to classify defects. The regression network estimates the depth of classified defects. A mild steel specimen with artificial delaminations is numerically modeled and excited by a quadratic frequency-modulated heat flux. The proposed network classification and regression performances are qualitatively assessed using testing time, accuracy, and mean squared error as a figure of merits.

The active thermal nondestructive testing and evaluation method is a rapidly growing testing procedure for a quick and remote inspection procedure for fibre‐reinforced plastics. Conventional modulated lock‐in thermography significantly contributed to this field by allowing usage of low peak power controlled stimulations followed by phase based detail extraction procedures. But demand of repetitive experimentation required for depth scanning of the test object limits its applicability for realistic critical applications and demands multi‐frequency low power stimulations for better resolution and sensitivity for sub‐surface defect detection. Frequency modulated thermal wave imaging and coded excitation thermal wave imaging methods permitting multi‐frequency stimulations cater for these needs and facilitate depth scanning of the test object in a single experimentation cycle. Recently introduced three‐dimensional pulse compression is an alternative to phase based analysis for these stimulations by providing enhanced defect detection even in noisy environmental and experimental conditions. Defect detection capability and sizing by these non‐stationary thermal wave imaging methods are highlighted using the pulse compression approach. The present experimental study has been carried out on a carbon fibre reinforced plastic specimen with flat bottom holes.

Purpose This study aims to at numerically retrieve five constant dimensional thermo-physical properties of a biological tissue from dimensionless boundary temperature measurements. Design/methodology/approach The thermal-wave model of bio-heat transfer is used as an appropriate model because of its realism in situations in which the heat flux is extremely high or low and imposed over a short duration of time. For the numerical discretization, an unconditionally stable finite difference scheme used as a direct solver is developed. The sensitivity coefficients of the dimensionless boundary temperature measurements with respect to five constant dimensionless parameters appearing in a non-dimensionalised version of the governing hyperbolic model are computed. The retrieval of those dimensionless parameters, from both exact and noisy measurements, is successfully achieved by using a minimization procedure based on the MATLAB optimization toolbox routine lsqnonlin. The values of the five-dimensional parameters are recovered by inverting a nonlinear system of algebraic equations connecting those parameters to the dimensionless parameters whose values have already been recovered. Findings Accurate and stable numerical solutions for the unknown thermo-physical properties of a biological tissue from dimensionless boundary temperature measurements are obtained using the proposed numerical procedure. Research limitations/implications The current investigation is limited to the retrieval of constant physical properties, but future work will investigate the reconstruction of the space-dependent blood perfusion coefficient. Practical implications As noise inherently present in practical measurements is inverted, the paper is of practical significance and models a real-world situation. Social implications The findings of the present paper are of considerable significance and interest to practitioners in the biomedical engineering and medical physics sectors. Originality/value In comparison to Alkhwaji et al. (2012), the novelty and contribution of this work are as follows: considering the more general and realistic thermal-wave model of bio-heat transfer, accounting for a relaxation time; allowing for the tissue to have a finite size; and reconstructing five thermally significant dimensional parameters.

Conventional infrared thermography techniques, including pulsed and lock-in thermography, have shown great potential for non-destructive evaluation of broad spectrum of materials, spanning from metals to polymers to biological tissues. However, performance of these techniques is often limited due to the diffuse nature of thermal wave fields, resulting in an inherent compromise between inspection depth and depth resolution. Recently, matched-filter thermography has been introduced as a means for overcoming this classic limitation to enable depth-resolved subsurface thermal imaging and improving axial/depth resolution. This paper reviews the basic principles and experimental results of matched-filter thermography: first, mathematical and signal processing concepts related to matched-fileting and pulse compression are discussed. Next, theoretical modeling of thermal-wave responses to matched-filter thermography using two categories of pulse compression techniques (linear frequency modulation and binary phase coding) are reviewed. Key experimental results from literature demonstrating the maintenance of axial resolution while inspecting deep into opaque and turbid media are also presented and discussed. Finally, the concept of thermal coherence tomography for deconvolution of thermal responses of axially superposed sources and creation of depth-selective images in a diffusion-wave field is reviewed.

Recent trends in thermal non-destructive testing focusing on artificial intelligence and various deep learning architectures have been investigated for quality assessment of different materials. The present work introduces three famous computer vision models (AlexNet, GoogleNet and VggNet) with one-dimensional convolution layers for defect detection for material inspected by quadratic frequency modulated thermal wave imaging. These models employ sequential convolution operations and pooling on temporal thermal profiles and extract deep features further to classify defect and sound regions in the test sample. The three deep learning models are trained from scratch with the experimental thermographic data of a carbon fiber reinforced polymer (CFRP) specimen with artificially simulated flat bottom hole defects of different sizes at varying depths. The performance metrics conclude that AlexNet presents high testing accuracy and F-score of 98.92% and 0.954 resulting in less deviation to the actual labels favoring enhanced defect signal-to-noise ratio with less computation time in CPU-based hardware. Further, the depth of the detected defect was quantified using a recently introduced quantification model using the chirp-z transform-based phase analysis. The estimated depths are rearranged in the respective locations and visualized the depth map.