Thermography Research Articles

Enhancing automated fiber placement process monitoring and quality inspection: A hybrid thermal vision based framework

Automated Fiber Placement (AFP) has revolutionized composite manufacturing, yet quality assurance remains challenging due to the significant impact of emerging defects on part quality and the current reliance on time-consuming manual inspection protocols. This paper presents a comprehensive hybrid framework that enhances AFP process monitoring and quality inspection by integrating thermal vision with innovative methodologies. Our framework combines model-based and data-driven algorithms across three modules to address key AFP inspection tasks, including in-situ monitoring, dynamic tow identification, defect detection, segmentation, localization, and quantitative lay-up quality evaluation. The setup-independent spatial–temporal analysis algorithm estimates tow boundaries with sub-pixel accuracy. An optimized SVM classifier, trained on an extensive AFP defect database, achieves a defect detection accuracy of 96.4% and an F1-score of 96.43%, meeting industry standards. The active contours-based segmentation and localization module provides critical qualitative traits such as defect shape, size, and location. Moreover, the novel Defect Area Percentage (DAP) metric enables precise quantitative defect impact evaluation at both the course and tow levels. By consolidating qualitative and quantitative outcomes, the system offers real-time high-level feedback for informed decision-making, significantly improving process performance and reducing machine downtimes. This proactive approach advances AFP process monitoring and quality inspection and positions our framework as a promising solution for next-generation composite manufacturing.

Robot-based dynamic optics scanning infrared thermography for debonding defect detection of big-sized CFRP specimen

ABSTRACT The use of infrared thermography (IRT) for non-destructive testing (NDT) of large aerospace composites, such as carbon fibre reinforced polymer (CFRP), is often limited by the narrow field of view infrared cameras. This paper experimentally investigates the big-sized CFRP specimens with simulated debonding defects using a robot-based dynamic optics scanning infrared thermography (RDOS-IRT) method to meet these issues. Then, static pulse IRT and RDOS-IRT were performed, and the processing performance of different algorithms (pulsed phase thermography, PPT; total harmonic distortion, THD; matched filtering, MF; and supervised dimensionality reduction, SDR) on pulsed IRT was discussed. A combination of pseudo-static matrix reconstruction (FSMR) algorithm and static image sequence processing algorithm is innovatively adopted to improve the defect detection performance of RDOS-IRT results. The experiments show that the RDOS-IRT method with FSMR and static image sequence processing algorithm is effective; moreover, it has almost a similar detection ability compared to the traditional static pulse IRT method in practical applications and saves more time. Results show that the signal-to-noise ratio (SNR) of the repeated defects is not differentiated, and the detection effect is similar. Therefore, the RDOS-IRT method is a reliable tool for NDT tasks on big-sized CFRP specimens and can be practically applied in an industrial setting for high detection efficiency.

Advancing mastitis assessment in dairy bovines via short milking tube thermography: A seasonal perspective.

In India, where dairy production leads globally, infrared thermography (IRT) and short milking tube thermography specifically are vital for managing mastitis. Therefore, the present study focuses on thermal imaging of the udder and short milking tube (SMT) of the milking machine during the peak milking process of Sahiwal cows and Murrah buffaloes during winter, summer, rainy and autumn seasons to identify sub-clinical (SCM) and clinical mastitis (CM) cases using the Darvi DTL007 camera. The udder health was assessed using the California Mastitis Test, Somatic Cell Count (SCC) and IRT throughout the year. Log10SCC and thermogram analysis revealed a difference (p < 0.01) between healthy, SCM, and CM cases during different seasons in both breeds. Further results showed an increase (p < 0.01) in SMT thermograms of SCM and CM cases compared to healthy quarters in Sahiwal cows during winter, summer, rainy, and autumn were 4.26 and 7.51, 2.37 and 4.47, 2.20 and 3.64, 2.90 and 4.94 ºC, respectively and for Murrah buffaloes were 3.56 and 5.55, 2.70 and 3.81, 1.72 and 3.10, 3.14 and 4.42ºC, respectively. The highest degree of increase in milking udder skin surface temperature and SMT of SCM and CM cases compared to healthy quarters was observed during the winter and the least during the rainy season. Thus, regardless of the seasons examined in this study, SMT thermograms effectively assessed SCM and CM.

Infrared thermography and computed tomography imaging for hind limb study after immobilization-induced disuse atrophy

Immobilization for treatment after an injury can lead to disuse atrophy, resulting in reduced functionality and strength of the immobilized limb. In our study, we utilized infrared thermography (IR) and computed tomography (CT) ex vivo to assess both physiological and structural changes following hind limb immobilization in a young Wistar rat model. Twelve rats weighing 275 ± 30 g had their right hind limbs immobilized with a modified Thomas-splint for varying durations (3, 7, or 14 days). IR imaging using an infrared camera provided insight into limb temperature changes. For micro-CT, we implemented a stain-ethanol fixation method and a gray score which enabled us to visualize and quantify muscle alterations. Thermographic images showed an increase in temperature of up to 8% in the hind limb at supine position at 14 days due to the inflammatory process while micro-CT exhibited muscle shrinkage of 10 and 18% at 7 and 14 days, respectively. Our findings underscore the efficacy of IR and micro-CT as rapid and precise imaging modalities for detecting morphological shifts in muscle tissue, particularly in pathological conditions like atrophy.

Code-specified early delamination detection and quantification in a RC bridge deck: passive vs. active infrared thermography

Code-specified early delamination detection and quantification in a RC bridge deck: passive vs. active infrared thermography

Seasonal assessment of mastitis in crossbred goats: A thermographic approach

Sub-clinical mastitis exhibits a higher prevalence in dairy goats than clinical mastitis, necessitating the adoption of non-invasive diagnostic techniques such as infrared thermography (IRT) to detect this economically significant production disease in the dairy sector. Accordingly, this study aims to employ IR imaging of the udder and teat quarters of lactating crossbred goats (Alpine × Beetal and Sanen × Beetal) across various seasons, utilising IRT, to discern cases of sub-clinical (SCM) and clinical mastitis (CM). Over a year, 100–110 lactating crossbred dairy goats underwent consistent IRT screenings, followed by a comprehensive evaluation of udder health status using the California mastitis test and somatic cell count (SCC). The receiver operating characteristic (ROC) analysis was conducted to establish the cut-off values for different thermographic parameters in this study. The results revealed that the SCC increased significantly (p < 0.01) in healthy, SCM, and CM milk samples across the seasons. The analysis of ROC revealed a comparatively higher sensitivity, specificity, and accuracy for udder thermograms during SCM than CM and vice versa for teat thermograms. IRT analysis reflected a difference (p < 0.01) in the udder and teat thermograms among quarters of healthy, SCM, and CM in summer, winter, autumn, and rainy seasons. A significant increase (p < 0.01) in udder thermograms was observed for quarters affected with SCM and CM relative to healthy, with an increase of 1.89 and 2.94 °C in winter, 0.85 and 1.63 °C in summer, 0.73 and 1.41 °C in rainy, and 1.33 and 2.38 °C in autumn, respectively. Similarly, for teat thermograms it was 1.79 and 2.81 °C in winter, 0.76 and 1.41 °C in summer, 0.70 and 1.37 °C in rainy, and 1.09 and 1.93 °C in autumn, respectively. Therefore, regardless of the seasons examined in this study, IRT proves to be an effective and supportive tool for early mastitis detection in lactating crossbred dairy goats.

ThermalTrend: a tool for organising and analysing thermal images

Infrared thermography (IRT), as a non-contact condition monitoring (CM) technique, measures the temperature distribution of machines and components to detect anomalous thermal behaviour. It is usually considered as a complementary technique and nowadays is used in a wide variety of industrial applications. In this paper, two commonly used criteria to assess the condition of machines based on the infrared thermography technique, ie absolute and relative temperature criteria, are represented. Then, the importance of the temperature trend in condition monitoring and fault diagnosis of machines and components is described. The ThermalTrend software, which was developed in the Condition Monitoring Unit, Smelting Department, Sarcheshmeh Copper Complex in Rafsanjan, Iran, is introduced and its most important and unique features are discussed. This software package helps thermographers to organise thermal images and provides the temperature trends of machines and components that are under regular monitoring. Finally, some industrial case studies on the usefulness and applicability of the infrared thermography technique and the ThermalTrend software are reported.

Fully automated unsupervised learning approach for thermal camera calibration and an accurate COVID-19 human temperature tracking

During the past three years, people have suffered greatly from what the World Health Organization called the emerging COVID-19. The world lacked the means and methods for early detection of this virus. Several traditional methods were used for detecting the virus, such as thermometers, remote thermal detection guns, and other conventional methods. Most of these systems monopolized making profits or selling their camera products, with prices for these cameras equipped with temperature detection systems exceeding three thousand dollars. An unsupervised model for real-time detection of thermal face skin temperature is proposed. Despite the scarcity and limited availability of thermal video data, we found and used a database created at Nazarbayev University in Nur-Sultan, Kazakhstan, which contains clips of thermal video and RGB video. The two different videos were calibrated, and the unity was measured by two metrics: SSIM and correlation. Four methods of registration were used to achieve perfect congruence, and congruence was also measured through the two previous metrics. The K-means method was then used to extract clusters, and functions for post-processing were built. The thermal face skin was extracted by multiplying the binary face mask with the thermal face, and the temperature of the face was calculated by taking the average values of the thermal face skin pixels and converting them from Fahrenheit to Celsius. Satisfactory results were obtained, with some temperatures detected within the normal range, others below the normal range, and others higher than this rate.

Compression–compression fatigue damage of wrinkled carbon/glass hybrid composite laminates

Compression–compression fatigue tests were carried out to study the compressive fatigue damage mechanisms of a carbon/glass hybrid composite laminate with a manufacturing-induced wrinkle defect. As reported in literature, thin wrinkled composite laminates could show sensibly different fatigue behaviours and damage mechanisms in the presence of tension and compression cyclic loadings. The sensitivities of composites to tension and compression cyclic loadings were not the same. In this study, the damage modes and cracking sequence of a thick wrinkled laminate were identified by the strain fields from digital image correlation (DIC). Acoustic emission (AE) and infrared (IR) thermography were used to detect the damages and explore the potential capabilities of non-destructive evaluation methods when applied to detection of fatigue induced damages in composite structures under operational cyclic loadings. Two primary damage modes were identified and detected before the final failure of the laminate i.e. debonding between the resin-rich layer and the laminate and interlaminar cracks. Debonding occurred earlier than interlaminar cracking. Out-of-plane stresses due to fibre waviness drove the initiation of these damages. Under the settings of AE in this work, AE captured early debonding activities but did not detect the micro damage accumulating before the formation of interlaminar crack. Interlaminar crack initiation was tracked by IR thermography at cross-section of the specimen based on the distinct temperature localisations that occur with this micro damage. No distinct temperature localisations occurred during the formation of debonding as it is mode-I driving under compression–compression cyclic loading, so IR thermography did not detect the debonding crack. The results of this study show the potential of using complementary non-destructive damage evaluation methods to inspect early damages in composite structures. The early detection of damages would give early warning signs before the damages become critical. It is found that AE and IR thermography should be used as complementary tools to detect different damage modes.

Assessing the performance of infrared thermography and PIV methods to identify and characterize the air inlets and outlets in wall assemblies

Building airtightness is crucial for reducing energy demand in the building sector and preventing moisture damage and indoor environment quality issues. Mandatory airtightness tests are now commonly performed in many countries to quantify the total envelope leakage rate of real buildings. For a more precise diagnosis, air leakages are usually identified by infrared thermography or visualized by smoke. However, a finer analysis is also needed to characterize precisely the air flow through specific leaks. Numerical models have been developed for this purpose but experimental data are necessary to validate them. In this paper, two experimental approaches are tested on simple light-wall assembly configurations to address this need: infrared thermography (IRT) and the innovative use of particle image velocimetry (PIV) with a specific experimental set-up and protocol developed. The results are compared with each other as well as with numerical simulations. Results showed that both techniques can be applied for the experimental characterization of the air inlet/outlet of a wall assembly with different advantages and drawbacks. IRT is easy to implement, non-intrusive, possible for in-situ investigations, but the correlation between the thermographs and the air dispersion at the outlet is not straight forward due to thermal inertia and heat conduction issues. PIV gives direct results on the velocity field and enables quantitative analysis of air flow rates for various configurations, with some limitations: a difficult implementation; restrictions on the possible infiltration velocity range and difficulties to visualize the flow exactly at the interface with the wall assembly.

Thermal nondestructive testing of cracks in turbine blades by using ultrasonic stimulation

The results of nondestructive testing (NDT) of a turbine blade made of a heat resistant alloy with a ceramic coating were obtained by using ultrasonic infrared (IR) thermography. The purpose of the study was to determine possibilities of this NDT technique in detecting cracks in the blade and the coating. Image processing was performed by using principal component analysis, which allows to underline defect indications in IR images. The obtained test results were in a good accordance with the results obtained by means of penetrants with a considerably shorter inspection time. Potentials of IR ultrasonic thermography in the detection of “kissing: cracks and cracks located in difficult to reach sites was demonstrated.

Are airports like cities? Affordances and people’s micro embodied interactions during the arrival experience

Airports are designed to work efficiently. Specifically, they seek to efficiently process and ‘control’ people’s movements across landside and airside spaces, as well as to entice consumption through the architectural engineering of atmospheres and passenger experiences using architecture (Adey 2008; Fuller and Harley 2004; Hubregtse 2020). However, studies of the design and human sensorial perception of spaces of mobilities have shown that the way people move through and inhabit such spaces transcends the intended functionality of their design (Jensen and Lanng 2017). This includes work that considers airports as not only machines of movement, control and profit, but also city-like spaces (Hernandez-Bueno 2021; Nikolaeva 2013). Based on the conceptualisation of the airport as an ambiguous and city-like space, this paper uses thermal camera video recordings, ethnography, and design mappings to analyse the passenger movements, practices, embodied interactions, and mobile affordances of Copenhagen Airport’s meet and greet area. Focusing on a micro scale, the analysis shows the comparability of people’s practices and ways of inhabiting the airport with practices common to urban public spaces. This facilitates broader reflection on how to speculatively re-design and re-imagine the airport like an urban ‘public’ space.

Explainable artificial intelligence prediction of defect characterization in composite materials

Non-destructive evaluation (NDE) techniques are integral across diverse applications for void detection within composites. Infrared (IR) thermography (IRT) is a prevalent NDE technique that utilizes reverse heat transfer principles to infer defect characteristics by analyzing temperature distribution. Although the forward heat transfer problem is well-posed, its inverse counterpart lacks uniqueness, posing non-unique solutions. The present study performs simulations using finite element analysis (FEA) in defective (a penny-shaped defect) composites through which the heat transfer flux is modeled. A total of 2100 simulations with various defect positions and sizes (depth, size, and thickness) are executed, and the corresponding surface temperature vs. time and vs. distance diagrams are extracted. The FEA outputs provide ample input data for developing an explainable artificial intelligence (XAI) model to estimate the defect characteristics. A detailed feature engineering task is performed to select the representative information from the diagrams. Explainable decision tree-based machine learning (ML) models with transparent decision paths based on derived features are developed to predict the defect depth, size, and thickness. The ML models’ results suggest superb accuracy (R2 = 0.92 to 0.99) across all three defect characteristics. The provided workflow sets a benchmark applicable to a range of fields, including health monitoring.

Hypervigilance to pain and sleep quality are confounding variables in the infrared thermography examination of the temporomandibular joint and temporal and masseter muscles.

Verify whether hypervigilance to pain (HP) and sleep quality (SQ) are confounding variables in the infrared thermography (IT) examination of the temporomandibular joint and temporal and masseter muscles. A cross-sectional and analytical study was conducted, collecting HP and SQ data from 80 participants without temporomandibular disorders (TMD), performing their IT and another 40 participants with TMD. For the selection of participants with and without TMD, the TMD Pain Screener questionnaire and axis I of the Diagnostic Criteria for Temporomandibular Disorders were applied. SQ was verified using the Pittsburgh Sleep Quality Index (PSQI) questionnaire. For the HP assessment the Pain Vigilance and Awareness Questionnaire (PVAQ) was applied. And the IT was performed through a FLIR infrared sensor camera, model T650 Infrared. No significant correlations were found between SQ and the temperatures of the areas of interest (P > .05), and regarding HP, a statistically significant positive correlation was found with the dimensionless (ρ = 0.289) and non-dimensionless (ρ = 0.223) asymmetries of temporal muscle temperatures. In the temperature comparisons between the participants without TMD and the participants with TMD, significant differences were found (P < .05), also when the group without TMD was controlled according to both HP and SQ (P < .05), with higher temperatures found in the TMD group. HP and SQ can be considered confounding variables in IT examination of the temporomandibular region.

Thermal profiles of Cicindelidia haemorrhagica (Coleoptera: Cicindelidae) activity in hot springs in Yellowstone National Park.

The wetsalts tiger beetle, Cicindelidia haemorrhagica (LeConte) (Coleoptera: Cicindelidae), is found in several active thermal hot spring areas in Yellowstone National Park (YNP) where substrate surface temperatures can exceed 50 °C. However, relationships between surface temperatures and the time adults spend on them remain poorly understood. Therefore, we characterized thermal profiles of Dragon Spring and Rabbit Creek, 2 thermally active research sites containing C. haemorrhagica in YNP, to quantify the time adults spend at different surface temperatures. We took 58 thermal video recordings of adults over 6 total days of observation ranging from 10 to 15 min for each adult. Thermal video analysis results indicated a positive relationship between the total time adult beetles spent on surface temperatures from Dragon Spring and Rabbit Creek as temperatures increased from 20 °C. Once surface temperatures exceeded 40 °C, the total time spent at those surface temperatures declined. Adults were recorded on substrates exceeding 50 °C at one of the 2 research locations. Rabbit Creek had substantially more instances of adults present with surface temperatures exceeding 40 °C, including one individual on a surface temperature of 61.5 °C. There were 3 instances of beetles spending more than 4 min at a particular surface temperature, all within the preferred range of 30-40 °C. Our thermal profile results and previous behavioral observations suggest that adults may be resistant to the heat produced from the thermal waters that influence the substrate temperatures but may not be subject to high surface temperatures as previously reported.

Thermal imaging analysis of ballast fouling: Investigating the effects of parent rock and fouling materials through IRT passive camera

This study explores the use of infrared thermography (IRT) technology for the non-destructive evaluation of ballast fouling in railway tracks, focusing on the influence of parent rock types and fouling materials. Utilizing thermal imaging, the research investigates how variations in ballast conditions affect surface temperature, which serves as an indicator of structural integrity and health. The experimental setup involved ballast samples derived from three different rock types—basalt, limestone, and andesite—fouled with commonly encountered materials like sand and clay at varying percentages. Results demonstrate that fouling level and type significantly influence the thermal signatures captured by IRT passive camera. Notably, ballast derived from darker rocks exhibited higher temperatures, indicating greater emissivity, while fouled ballast showed distinct temperature patterns compared to clean samples, emphasizing the potential of thermal imaging in detecting and quantifying fouling in ballast layers. This research underscores the viability of IRT passive camera in the routine maintenance and monitoring of railway infrastructure, providing a foundation for further development of integrated diagnostic tools for railway management systems.

Thermal preference prediction through infrared thermography technology: Recognizing adaptive behaviors

The recognition of thermal adaptation behaviors is a crucial and noninvasive method for accurately and swiftly predicting the thermal preferences of occupants. This is instrumental in enhancing the energy efficiency, indoor personnel comfort, and overall productivity of office buildings. This study employed deep-learning models to analyze a thermal adaptive behavior video dataset captured by Infrared thermography (IRT) cameras to predict the thermal preferences of occupants. A large-scale infrared thermography dataset was developed for recognizing thermal adaptive behaviors using 9650 video samples. Two adaptive behavior recognition models for infrared videos were developed by leveraging neural network models: Two-Stream Inflated 3D ConvNet (I3D) and SlowFast (SF) networks. These models achieved the average prediction accuracies of 83.53 % and 88.56 %, respectively. Notably, both models exhibited a recognition time of milliseconds, facilitating the real-time recognition of thermal adaptive behaviors. This study offers vital technical and theoretical underpinnings for developing decision-making solutions for smart buildings based on noninvasive thermal preference prediction.

An Experimental Investigation Of Local Thermo-Fluidic Heat Transfer In A U-Shaped Mini-Channel Sink

This study investigates experimentally a heated U-shaped mini-channel heat sink using Infrared Thermography and Particle Image Velocimetry for a water coolant flow of Reynolds numbers of 280, 650, and 1350 (based on the hydraulic channel diameter). The choice of this cell geometry is based on its role as a simplified unit of a serpentine heat exchanger, which is proved to be one of the most promising for cooling processes. The use of the infrared camera allows the detection of temperature fields on top of the external surface of the cell. Therefore, aiming to derive the temperature distribution on the channel roof, a dedicated transfer function is implemented. Moreover, we employed the lumped capacitance model for thermal analysis on both infrared measurements and thermocouple data. The latter are recorded to capture the cooling process of the aluminium base and water temperature at the end of the outlet tube. As a result, thermal transient rate, cooling magnitude and equilibrium temperature are obtained. These parameters indicate that higher Reynolds numbers correspond to increased thermal transient rates, enhanced cooling effects, and lower equilibrium temperatures. A non-uniform distribution of heat transfer along the channel is reported, with the most efficient cooling area localized close to the first 90-degree corner. These findings are consistent with numerical simulations and previous experimental observations. PIV results reveal the presence of two fluid acceleration zones following both 90-degree corners, which contribute to improve the water cooling ability in their respective regions. Additionally, the formation of two recirculating bubbles is reported at the inner wall from corners vertices, whose intensity is dependent on Reynolds number, pushing the main flow towards the outer channel wall and reducing the local heat transfer. Turbulent kinetic energy distributions are also investigated, pointing out the presence of intense areas that better match with regions of minimum equilibrium temperature than maximum velocity zones. This suggests the presence of local turbulent unsteadiness and three-dimensional phenomena, which contributes significantly to cooling enhancement.

Thermomechanical effects during air blast freezing of fruit pulp model solutions with different sucrose concentrations in metal drums

The strains on the metal drums surfaces caused by the ice volumetric expansion during the freezing of model solutions of fruit pulps were evaluated by extensometers. 200 kg of model solutions of 0.5% k-carrageenan (w/v) added 0.5%, 10% or 20% (w/v) sucrose into metal drums were frozen in air blast freezer at −25 °C. Strain gauge sensors were installed on the side surface and bottom of the drums in a balanced Wheatstone bridge circuit. The resulting stresses and temperature profiles inside the packages were measured simultaneously. The non-uniform temperature distribution in the product was proven by IR thermography. There was reduction in the freezing temperature and stresses in the packages with the increase in the sucrose concentration. The restricted surfaces, preventing the expansion of the ice, yielded great stresses in the frozen solutions and induced cracking in the free surfaces of the solutions, especially in the systems with lower concentrations.

DBHF: A double fusion backbone and bidirectional feature hybrid fusion detector for high-precision inspection of avionics solder joint defects

The quality of avionics thermistor wire solder joints has a critical impact on the performance of in-orbit spacecraft. However, it is still hard to simultaneously realize a high-precision inspection of the multi-scale internal and external defects of avionics solder joints. Herein, in combination with active infrared thermography technique, an innovative DBHF detector with a double fusion backbone and a learnable weighted multi-scale bidirectional feature hybrid fusion pyramid network (BdFPN) is demonstrated based on RT-DETR framework. Specifically, the double fusion backbone framework comprises with two lightweight convolutional neural network-based backbones. Compared with conventional single backbone, the double-backbone design forms a multi-level Resnet-like structure, helping to enrich the extracted multi-dimension feature information and retain rich detail information of small-scale defects. The complex learnable weighted multi-scale BdFPN is designed for cross fusion of deep features. One or more additional multi-scale inputs are added to each node involved in feature fusion to aggregate more features from multiple scales and strengthen the interaction between low-level details and high-level semantic information, enabling the detector to make full use of multi-level feature information. Meanwhile, a learnable weight coefficient is set on each node involved in feature fusion and the importance of the input to each node is estimated by training, so that the effective information can be deeply aggregated at different levels. Our DBHF detector achieves a 92.7 % mAP@0.5 on a self-made avionics solder joint dataset with a body size of only 27.2 MB and FPS of 149, outperforming all the typical advanced real-time detectors.