Multi-label classification algorithms for composite materials under infrared thermography testing

Alhammad, Muflih; Avdelidis, Nicolas P.; Ibarra-Castanedo, Clemente; Maldague, Xavier; Zolotas, Argyrios; Torbali, Ebubekir; Genest, Marc

doi:10.1080/17686733.2022.2126638

Cited by 8 publications

(3 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…IR thermography has been successfully used to study these underlying defects. 82 The technique involves of heating one part of the sample with high power laser and obtaining the thermal response on the other part as the heat is transferred through the sample. Samples with high porosity temperature rise showed lesser rise in temperature with longer time to reach the other fact.…”

Section: Optical Ndtmentioning

confidence: 99%

Damage detection in composites using non-destructive testing aided by ANN technique: A review

Saha

Roy

Topdar

2023

Journal of Thermoplastic Composite Materials

View full text Add to dashboard Cite

Damages are inevitable in structures and effective damage detection techniques are important for maintaining their health. Many weight-sensitive engineering applications use composite materials, especially fiber-reinforced laminates. Common damages of these materials include delamination, fiber breakage, fiber pull-out, etc. Various non-destructive testing (NDT) techniques are reported in the literature for damage detection in composites, such as ultrasonic testing, vibration-based techniques, acoustic emission technique, optical NDT and imagining techniques. However, due to the complex properties of composite materials, conventional techniques for analyzing NDT data are difficult to implement. In this context, artificial neural network (ANN) technique is a promising alternative for analyzing NDT data for damage detection. In this study, an attempt is made to explore the state-of-the-art of damage detection in composites using NDT aided by ANN. The work discusses the pros and cons of different methods and is expected to help in identifying the appropriate method for damage detection in composites.

show abstract

Section: Optical Ndtmentioning

confidence: 99%

Damage detection in composites using non-destructive testing aided by ANN technique: A review

Saha

Roy

Topdar

2023

Journal of Thermoplastic Composite Materials

View full text Add to dashboard Cite

show abstract

“…There are numerous approaches to tackle multi-label classification issues [13][14][15] and these approaches can be categorized as algorithm adaptation method and problem transformation method [16]. The first category includes algorithms like Rank-SVM [17], ML-DT [18], and ML-KNN [19], which can be extended to process multi-label data.…”

Section: Introductionmentioning

confidence: 99%

An Artificial Intelligence-Based Stacked Ensemble Approach for Prediction of Protein Subcellular Localization in Confocal Microscopy Images

Aggarwal

Gupta

et al. 2023

Sustainability

View full text Add to dashboard Cite

Predicting subcellular protein localization has become a popular topic due to its utility in understanding disease mechanisms and developing innovative drugs. With the rapid advancement of automated microscopic imaging technology, approaches using bio-images for protein subcellular localization have gained a lot of interest. The Human Protein Atlas (HPA) project is a macro-initiative that aims to map the human proteome utilizing antibody-based proteomics and related c. Millions of images have been tagged with single or multiple labels in the HPA database. However, fewer techniques for predicting the location of proteins have been devised, with the majority of them relying on automatic single-label classification. As a result, there is a need for an automatic and sustainable system capable of multi-label classification of the HPA database. Deep learning presents a potential option for automatic labeling of protein’s subcellular localization, given the vast image number generated by high-content microscopy and the fact that manual labeling is both time-consuming and error-prone. Hence, this research aims to use an ensemble technique for the improvement in the performance of existing state-of-art convolutional neural networks and pretrained models were applied; finally, a stacked ensemble-based deep learning model was presented, which delivers a more reliable and robust classifier. The F1-score, precision, and recall have been used for the evaluation of the proposed model’s efficiency. In addition, a comparison of existing deep learning approaches has been conducted with respect to the proposed method. The results show the proposed ensemble strategy performed exponentially well on the multi-label classification of Human Protein Atlas images, with recall, precision, and F1-score of 0.70, 0.72, and 0.71, respectively.

show abstract

“…Similarly, applications of step-heating thermography (SHT) on composite structures [ 23 , 24 , 25 , 26 ] are described as an alternative to pulsed thermography (PT) [ 19 , 27 , 28 , 29 ] for overcoming its limitations, as it is more applicable to thinner materials, and as an alternative to LT to overcome the limitations of long inspection time [ 23 ]. Apart from using heat as an excitation method in active thermography, Szymanik et al [ 30 ] suggest the use of forced cooling as an additional excitation method that can affect heat dissipation within a material (in their case a glass fibre composite) and can cause rapid temperature change in defective areas which can be detected by an infrared camera.…”

Section: Introductionmentioning

confidence: 99%

Exploring Step-Heating and Lock-In Thermography NDT Using One-Sided Inspection on Low-Emissivity Composite Structures for New Rail Carbodies

Tromaras

Kappatos

2022

Sensors

View full text Add to dashboard Cite

This paper aims to explore the qualification of step- and lock-in heating thermography as techniques capable of inspecting new composite rail carbodies following input and inspection requirements set by the rail manufacturing industry. Specifically, we studied (a) a monolithic CFRP sample (20 mm thickness) and (b) a CFRP–PET foam–CFRP sandwich (40 mm total thickness) component, that were manufactured with artificial defects, to replicate the side wall sections of a carbody. The samples proved to be very challenging to test using only one-sided inspection due to (1) exhibiting significant thickness compared to existing literature, (2) low surface emissivity and (3) that the foam core of the sandwich sample was a thermal insulating material. In addition, the sandwich sample was designed with defects on both skins. Both thermography techniques provided similar defect detection results, although step heating offered faster detection. In the case of the monolithic panel, defects up to 10 mm depth were detected, with minor detection of defects at 15 mm depth with a step-heating protocol between 90 s and 120 s overall acquisition, which was faster than the 140 s used with the lock-in technique. For the sandwich component only the front skin defects were detected, with both techniques using heating protocols between 70–120 s.

show abstract

Multi-label classification algorithms for composite materials under infrared thermography testing

Cited by 8 publications

References 37 publications

Damage detection in composites using non-destructive testing aided by ANN technique: A review

Damage detection in composites using non-destructive testing aided by ANN technique: A review

An Artificial Intelligence-Based Stacked Ensemble Approach for Prediction of Protein Subcellular Localization in Confocal Microscopy Images

Exploring Step-Heating and Lock-In Thermography NDT Using One-Sided Inspection on Low-Emissivity Composite Structures for New Rail Carbodies

Contact Info

Product

Resources

About