Filtern
Dokumenttyp
- Zeitschriftenartikel (9)
- Beitrag zu einem Tagungsband (9)
- Vortrag (5)
- Posterpräsentation (5)
- Preprint (3)
Schlagworte
- Super resolution (18)
- Laser thermography (16)
- Thermography (12)
- Joint sparsity (8)
- VCSEL array (7)
- Virtual wave (7)
- 1d laser (6)
- High-power laser (6)
- Laser array (6)
- NDT (5)
Organisationseinheit der BAM
Paper des Monats
- ja (1)
Eingeladener Vortrag
- nein (5)
Classification of Spot-Welded Joints in Laser Thermography Data Using Convolutional Neural Networks
(2021)
Spot welding is a crucial process step in various industries. However, classification of spot welding quality is still a tedious process due to the complexity and sensitivity of the test material, which drain conventional approaches to its limits. In this article, we propose an approach for quality inspection of spot weldings using images from laser thermography data. We propose data preparation approaches based on the underlying physics of spot-welded joints, heated with pulsed laser thermography by analyzing the intensity over time and derive dedicated data filters to generate training datasets. Subsequently, we utilize convolutional neural networks to classify weld quality and compare the performance of different models against each other. We achieve competitive results in terms of classifying the different welding quality classes compared to traditional approaches, reaching an accuracy of more than 95 percent. Finally, we explore the effect of different augmentation methods.
Learned block iterative shrinkage thresholding algorithm for photothermal super resolution imaging
(2020)
Block-sparse regularization is already well-known in active thermal imaging and is used for multiple measurement based inverse problems. The main bottleneck of this method is the choice of regularization parameters which differs for each experiment. To avoid time-consuming manually selected regularization parameters, we propose a learned block-sparse optimization approach using an iterative algorithm unfolded into a deep neural network. More precisely, we show the benefits of using a learned block iterative shrinkage thresholding algorithm that is able to learn the choice of regularization parameters. In addition, this algorithm enables the determination of a suitable weight matrix to solve the underlying inverse problem. Therefore, in this paper we present the algorithm and compare it with state of the art block iterative shrinkage thresholding using synthetically generated test data and experimental test data from active thermography for defect reconstruction. Our results show that the use of the learned block-sparse optimization approach provides smaller normalized mean square errors for a small fixed number of iterations than without learning. Thus, this new approach allows to improve the convergence speed and only needs a few iterations to generate accurate defect reconstruction in photothermal super resolution imaging.
This paper presents deep unfolding neural networks to handle inverse problems in photothermal radiometry enabling super resolution (SR) imaging. Photothermal imaging is a well-known technique in active thermography for nondestructive inspection of defects in materials such as metals or composites. A grand challenge of active thermography is to overcome the spatial resolution limitation imposed by heat diffusion in order to accurately resolve each defect. The photothermal SR approach enables to extract high-frequency spatial components based on the deconvolution with the thermal point spread function. However, stable deconvolution can only be achieved by using the sparse structure of defect patterns, which often requires tedious, hand-crafted tuning of hyperparameters and results in computationally intensive algorithms. On this account, Photothermal-SR-Net is proposed in this paper, which performs deconvolution by deep unfolding considering the underlying physics. This enables to super resolve 2D thermal images for nondestructive testing with a substantially improved convergence rate. Since defects appear sparsely in materials, Photothermal-SR-Net applies trained blocksparsity thresholding to the acquired thermal images in each convolutional layer. The performance of the proposed approach is evaluated and discussed using various deep unfolding and thresholding approaches applied to 2D thermal images. Subsequently, studies are conducted on how to increase the reconstruction quality and the computational performance of Photothermal-SR-Net is evaluated.
Thereby, it was found that the computing time for creating high-resolution images could be significantly reduced without decreasing the reconstruction quality by using pixel binning as a preprocessing step.
The separation of two closely spaced defects in fields of Thermographic NDE is very challenging. The diffusive nature of thermal waves leads to a fundamental limitation in spatial resolution. Therefore, super resolution image reconstruction can be used. A new concerted ansatz based on spatially structured heating and joint sparsity of the signal ensemble allows for an improved reconstruction of closely spaced defects. This new technique has been studied using a 1D laser array with randomly chosen illumination pattern.
This paper presents the results after applying super resolution algorithms, such as the iterative joint sparsity (IJOSP) algorithm, to our processed measurement data. Different data processing techniques before applying the IJOSP algorithm as well as the influence of regularization parameters in the data processing techniques are discussed. Moreover, the degradation of super resolution reconstruction goodness by the choice of experimental parameters such as laser line width or number of measurements is shown.
The application of the super resolution results in a spatial resolution enhancement of approximately a factor of four which leads to a better separation of two closely spaced defects.
In this paper we propose super resolution measurement and post-processing strategies that can be applied in thermography using laser line scanning. The implementation of these techniques facilitates the separation of two closely spaced defects and avoids the expected deterioration of spatial resolution due to heat diffusion. The experimental studies were performed using a high-power laser as heat source in combination with pulsed thermography measurements (step scanning) or with continuous heating measurements (continuous scanning). Our work shows that laser line step scanning as well as continuous scanning both can be used within our developed super resolution (SR) techniques. Our SR techniques make use of a compressed sensing based algorithm in post- processing, the so-called iterative joint sparsity (IJOSP) approach. The IJOSP method benefits from both - the sparse nature of defects in space as well as from the similarity of each measurement. In addition, we show further methods to improve the reconstruction quality e.g. by simple manipulations in thermal image processing such as by considering the effect of the scanning motion or by using different optimization algorithms within the IJOSP approach. These super resolution image processing methods are discussed so that the advantages and disadvantages of each method can be extracted. Our contribution thus provides new approaches for the implementation of super resolution techniques in laser line scanning thermography and informs about which experimental and post-processing parameters should be chosen to better separate two closely spaced defects.
Optical lock-in thermography is a completely contactless and very sensitive NDT technique. As an optical source of energy, incandescent lamps are most commonly used because they are relatively inexpensive and offer high irradiances at the test specimen. However, they are strongly restricted by their low modulation bandwidth with a maximum modulation frequency of only about 1 Hz. The use of high-power kilowattclass laser sources, e.g. diode laser arrays, pushes this constraint beyond 100 Hz. This allows for the exploration of the near-surface region of metals and layer systems with better and more accurate penetration depth and depth resolution. Moreover, these lasers are virtually free of any additional thermal radiation that could interfere with the “true” thermal response emitted from the heated sample. In turn, they can be easily used in a one-sided test configuration. We present current activities with kilowatt-class highpower laser sources for advanced lock-in thermography and focus on the application of laser arrays that offer a very high irradiation strength over a large sample area beyond the mentioned advantages.
This paper presents different super resolution reconstruction techniques to overcome the spatial resolution limits in thermography. Pseudo-random blind structured illumination from a onedimensional laser array is used as heat source for super resolution thermography. Pulsed thermography measurements using an infrared camera with a high frame rate sampling lead to a huge amount of data. To handle this large data set, thermographic reconstruction techniques are an essential step of the overall reconstruction process. Four different thermographic reconstruction techniques are analyzed based on the Fourier transform amplitude, principal component analysis, virtual wave reconstruction and the maximum thermogram. The application of those methods results in a sparse basis representation of the measured data and serves as input for a compressed sensing based algorithm called iterative joint sparsity (IJOSP). Since the thermographic reconstruction techniques have a high influence on the result of the IJOSP algorithm, this paper Highlights their Advantages and disadvantages.
The separation of two closely located defects in fields of Thermographic NDE is very challenging. The diffusive nature of thermal waves leads to a fundamental limitation in spatial resolution. Therefore, super resolution image reconstruction can be used.
The measured thermal waves can be transformed to virtual (ultrasound) waves that can be processed by applying ultrasound reconstruction algorithms and finally the super resolution algorithm. Otherwise, it is also possible to make use of a Fourier transform with a subsequent super resolution routine.
These super resolution thermographic image reconstruction techniques in post-processing are discussed and evaluated regarding performance, accuracy and repeatability.
Learned block iterative shrinkage thresholding algorithm for photothermal super resolution imaging
(2022)
Block-sparse regularization is already well known in active thermal imaging and is used for multiple-measurement-based inverse problems. The main bottleneck of this method is the choice of regularization parameters which differs for each experiment. We show the benefits of using a learned block iterative shrinkage thresholding algorithm (LBISTA) that is able to learn the choice of regularization parameters, without the need to manually select them. In addition, LBISTA enables the determination of a suitable weight matrix to solve the underlying inverse problem. Therefore, in this paper we present LBISTA and compare it with state-of-the-art block iterative shrinkage thresholding using synthetically generated and experimental test data from active thermography for defect reconstruction. Our results show that the use of the learned block-sparse optimization approach provides smaller normalized mean square errors for a small fixed number of iterations. Thus, this allows us to improve the convergence speed and only needs a few iterations to generate accurate defect reconstruction in photothermal super-resolution imaging.
This article presents deep unfolding neural networks to handle inverse problems in photothermal radiometry enabling super-resolution (SR) imaging. The photothermal SR approach is a well-known technique to overcome the spatial resolution limitation in photothermal imaging by extracting high-frequency spatial components based on the deconvolution with the thermal point spread function (PSF). However, stable deconvolution can only be achieved by using the sparse structure of defect patterns, which often requires tedious, handcrafted tuning of hyperparameters and results in computationally intensive algorithms. On this account, this article proposes Photothermal-SR-Net, which performs deconvolution by deep unfolding considering the underlying physics. Since defects appear sparsely in materials, our approach includes trained block-sparsity thresholding in each convolutional layer. This enables to super-resolve 2-D thermal images for nondestructive testing (NDT) with a substantially improved convergence rate compared to classic approaches. The performance of the proposed approach is evaluated on various deep unfolding and thresholding approaches. Furthermore, we explored how to increase the reconstruction quality and the computational performance. Thereby, it was found that the computing time for creating high-resolution images could be significantly reduced without decreasing the reconstruction quality by using pixel binning as a preprocessing step.