PSI - Issue 77

Andrzej Katunin et al. / Procedia Structural Integrity 77 (2026) 18–25 Author name / Structural Integrity Procedia 00 (2026) 000–000

19

2

1. Introduction Thermographic assessment of composite structures found wide applicability in structural inspections, particularly for the elements of aircraft components, being one of the FAA certified non-destructive techniques (FAA, 2014). The inspections are performed using various specific techniques, such as transient thermography, flash thermography, pulsed thermography, lock-in thermography, ultrasound vibrothermography, and others (see (Ibarra-Castanedo et al., 2013) for more information), which demonstrate applicability in various testing conditions and for various inspection purposes. Most of the mentioned techniques are based on external thermal excitation using various sources, like halogen lamps, lasers, infrared radiators, hot air or even LEDs (Lizaranzu et al., 2015; Yang and He, 2016; Ciampa et al., 2018; Dahlberg et al., 2022). In 2018, the self-heating based vibrothermography (SHVT) was introduced (Katunin, 2018a), which allowed performing thermographic inspections without external heat source. The SHVT technique is based on the concept of transient thermography, however, the thermal excitation is performed through a conversion of mechanical energy from forced vibrations to thermal energy due to the appearance of the self-heating effect, resulting from viscoelastic behavior of polymers and polymer-matrix composites (PCMs). This technique was patented (Katunin, 2019) and successfully tested for various types of structural damage characteristic for PCMs (Katunin and Wachla, 2018; Katunin et al., 2019). The idea behind the effective damage identification using SHVT is an excitation of a tested structure by multi-harmonic signal composed of several natural frequencies of vibration of this structure. This ensures proper thermal excitation on the whole surface of a tested structure, since the amount of dissipated heat is proportional to the introduced stress, which is the highest for the locations with the highest vibration magnitudes within the excited mode shape. Therefore, using a single natural frequency does not allow acquiring thermal response in the nodes of a corresponding mode shape, and therefore, excitation a tested structure on several natural frequencies is necessary to omit the nodes in mode shapes. Recent attempts of the authors’ team (Amraei et al., 2025) resulted in extension of SHVT to 2D plate-like structures demonstrating effectiveness in identifying artificially introduced damage in PMC structures. One of the greatest challenges during assessment of a structural damage is the selection of a proper thermographic image, since the collected images in a sequence differ because of the non-stationary heat transfer. Several approaches to the selection of the best thermographic have been developed that can be applied to solve this problem. One of the popular approaches is based on signal-to-background contrast used, e.g., in (Ricci et al., 2024; Amraei et al., 2025). Another approach is based on using thermal signal reconstruction algorithm and determine the peak response from the acquires results, as suggested in (Balageas et al., 2015). The comparative analysis of various algorithms has been presented in (D’Accardi et al., 2018), where the authors studies the advantages and disadvantages of various approaches in solving this problem. The effective solution for the selection of the best thermogram from a sequence for SHVT was presented in (Amraei et al., 2025). Additionally, for quantification of damage from the selected thermographic image, it is useful to apply a dedicated image processing algorithm, which makes it possible to significantly enhance the detectability as well as characteristic geometrical properties of a damage. Such an algorithm should resolve numerous challenges present during processing of thermographic images, such as measurement noise, low contrast between health and damaged regions, edge blurring, and other artifacts. The enhancement algorithms used for thermograms cover a great variety of approaches, starting with simple estimation of simple statistical features, and ending with advanced transforms and algorithms, like Fourier and wavelet transforms, principal component thermography, thermographic signal reconstruction (Montanini et al., 2016; Chrysafi et al., 2017; Wang et al., 2018; Wu et al., 2018), and their modifications (Yousefi et al., 2017; Schager et al., 2020). A useful case-study using various approaches to enhancement of thermograms has been presented in (Panella et al., 2021). Comprehensive comparative studies of various enhancement algorithms for thermograms acquired during testing with SHVT technique (Wronkowicz et al., 2019; Katunin et al., 2019) demonstrated that complex processing algorithms do not necessarily provide the best enhancement, which was also observed during the newest studies (Amraei et al., 2025). The aim of this study is to demonstrate the new processing algorithm for enhancement and quantification of a structural damage in PMCs from thermograms acquired using the SHVT technique. The developed approach is based on analysis of a selected sequence of thermograms from the whole sequence of collected thermographic images for further processing, and a two-step algorithm that considers filtering and wavelet-based fusion to achieve the image for further quantification of a detected damage. The studies were performed on sequences of thermograms acquired