Historically, researchers and engineers have faced the challenge of choosing between recording and analyzing visible-light or thermal data during testing. While visible-light imaging provides structural and contextual information, thermal data reveals heat signatures and thermal patterns. Attempting to capture both sets of data and manually align them spatially—or even more challenging, synchronize them temporally—has proven to be both inconsistent and time-consuming, often yielding more questions than answers.

The FLIR Multispectral Imaging Xperience (MIX) introduces a groundbreaking solution, empowering researchers and engineers to seamlessly analyze, interpret, and share complex thermal data with enhanced accuracy. By effortlessly blending thermal and visible-light imagery, the system captures and synchronizes high-speed data at over 1,000 frames per second. This integration ensures a single, cohesive dataset with precise spatial and temporal alignment—eliminating missed details and reducing uncertainties, ultimately providing comprehensive insights into fast-moving phenomena.

The presentation will outline the technical details of the FLIR MIX system and showcase how these purpose-built solutions are designed to synchronize high-speed thermal and visible-light imagery with precision, delivering the insights researchers need to push the boundaries of discovery.

Long-span steel bridges are coated with heavy-duty anticorrosion coatings consisting of several layers of paint to prevent the corrosion of steel materials. When maintaining heavy-duty corrosion-resistant paint, the early detection of the loss of the top coat and exposure of the undercoat is important because deterioration progresses rapidly if the top coat is lost. However, the current maintenance method relies on visual inspection; therefore, more efficient and accurate detection methods need to be developed. Furthermore, paints of similar colors are often used for the top and intermediate coats for color matching; therefore, it is difficult to distinguish them in the visible range. Therefore, we focused on the differences in the near-infrared spectral characteristics between the top coat with a weather-resistant paint, such as a fluorine resin paint, and the intermediate coat with an epoxy resin paint. To develop a method for evaluating the deterioration of heavy-duty anticorrosion coatings, we conducted an experimental study on an actual long-span bridge. First, we employed a near-infrared hyperspectral imaging system to obtain the spectral characteristics of the top and intermediate coats and identified the wavelength range where the spectral characteristics of each paint differed. Next, we employed a near-infrared camera attached to a band-pass filter that transmitted only the wavelength range and cut other wavelength ranges, and we obtained near-infrared images. Consequently, we were able to confirm the exposure of the intermediate coat using near-infrared measurements.

Carbon fiber-reinforced polymer (CFRP) composites exhibit higher strength in the longitudinal direction than in the transverse direction of the fibers, making their fiber orientation evaluation crucial. In this study, a method for evaluating fiber orientation using halogen spot periodic heating and a non-lock-in type infrared camera is employed and applied to a unidirectional carbon fiber composite—a CFRP composite with unidirectional continuous fibers. Experimental results show that the in-plane thermal diffusivity in the fiber direction is significantly higher than in other directions. Therefore, the employed method successfully evaluates fiber orientation in the unidirectional carbon fiber composite.

In this research, we developed and tested an automated defect detection system for wind turbine blades using infrared thermography (IRT) and a deep learning model U-Net.

A feasibility study was conducted on the detection of corroded rebars inside concrete using sub-THz waves. In this study, spectral measurement were performed with a reflection type system in the 20-50 GHz frequency range. Measurements were conducted on steel plates corroded in salt water and on concrete specimens containing these corroded plates. The results confirmed that reflectance decrease as corrosion progress. Furthermore, it was demonstrated that the presence or absence of internal steel plates and corrosion can be detected up to a cover thickness of 20 mm. The frequency spectral peaks and their periodicity also provided a means to estimate the cover thickness.

Deep learning techniques are increasingly being incorporated into the inspection and maintenance of social infrastructure. In this study, we show that supervised deep learning applied to imaging data obtained sub-THz wave, the average recall exceeded 80% for all cover thicknesses of steel plate inside concrete, and more than 90% for rebar inside concrete with cover thickness up to 20 mm. Unsupervised deep learning enabled the classification for both steel plate and rebar, even at large cover thickness. These results are expected to improve the exploration depth, which has been limited in previous studies.

To detect delamination and internal void defects within sandwich composite ma-terials, such as those used in wind turbine blades, this study employs a Remote Impact Test (RIT), analyzing the dispersion characteristics of the generated stress waves. RITs were conducted on specimens that varied in both thickness and defect type. Time-frequency spectrograms and dispersion curves were then obtained using two time-frequency analysis techniques: wavelet analysis and reassigned spectrograms (derived from Short-Time Fourier Transformation). The accuracy of defect identifica-tion is demonstrably improved through the cross-examination of the findings from these methods.

The application of thermal barrier coatings (TBCs) for protecting mechanical components is widespread, particularly in high-temperature environments such as gas turbines and aero-engines. Ensuring the integrity of these coatings throughout their service life is essential, as their degradation can lead to delamination, ultimately compromising the underlying component. It has been demonstrated that monitoring the thermal diffu-sivity value along time allows the monitoring of degradation of the coatings. Common thermographic techniques like pulsed and lock-in thermography have been used so far. However, to enhance both the signal-to-noise ratio (SNR) and the accuracy of thermal property measurements, new active thermography techniques have been developed. These methods rely on optimized excitation schemes combined with advanced signal processing strategies. In this work, we first introduce the pulse-compression thermography approach, which employs pseudo-noise modulated excitation to monitor and estimate the thermal diffusivity of the coating layers.

In recent years, electromagnetic waves (terahertz waves) with frequencies between 0.1 and 10 THz, which exist between radio waves and light waves, have attracted much attention. These electromagnetic waves have both the linearity of light waves and the transparency of radio waves and are expected to be applied to the field of human non-destructive testing. While it is known that terahertz waves can be used to detect foreign matter inside an object, we thought that by irradiating terahertz waves to the object to be measured from various directions, it would be possible to analyze the location and direction of contamination by comparing the scattering of the terahertz waves irradiated to the foreign matter. The samples were biomass resources in a jar with an opening of 53 mm and a diameter of 66.8 mm, and an aluminum plate with 76 × 50 mm. When terahertz waves were irradiated from the side of the jar with the biomass resources in it and the aluminum plate inserted, the transmission was higher when the metal plate was parallel to the light source and detector. This indicates that the transmission tendency of terahertz waves changes depending on the position and angle of the metal strip inside with respect to the direction of terahertz wave irradiation. This transmission tendency enables us to locate the position of a foreign object by irradiating terahertz waves from multiple directions, which is expected to be applied not only to the removal of foreign objects but also to various non-destructive inspection.

Ground Penetrating Radar is a non-destructive tool for detecting subsurface structures. However, traditional image interpretation is often limited by medium complexity and noise. To improve detection efficiency and accuracy, this study combines deep learning techniques to develop an automatic embankment cavity identification system based on the YOLOv10 model.The research first constructs a training dataset containing GPR images of embankment cavities and expands the dataset through data augmentation strategies to enhance model adaptability. Subsequently, cross-validation is employed to fine-tune the hyperparameters of the YOLOv10 model, seeking optimal performance. The experimental results demonstrate that the YOLOv10 model successfully identifies cavities in radar images, achieving accuracy rates of nearly 90% and 97%. This study proves the potential of deep learning in GPR image analysis, effectively improving detection efficiency, accuracy, and automation levels, providing more reliable technical support for embankment safety inspection.