Infrared Imaging Research Articles

Modeling Thermal Infrared Image Degradation and Real-World Super-Resolution Under Background Thermal Noise and Streak Interference

Modeling Thermal Infrared Image Degradation and Real-World Super-Resolution Under Background Thermal Noise and Streak Interference

Breast cancer diagnosis using optimized deep convolutional neural network based on transfer learning technique and improved Coati optimization algorithm

Breast cancer is a significant health concern due to its aggressive nature and high mortality rates. Early detection is crucial to improving patient outcomes. Thermography, a non-invasive and cost-effective method, utilizes heat from the breast surface to detect abnormalities, with the Database for Mastology Research (DMR) leveraging infrared images for research purposes. Deep learning (DL), particularly convolutional neural networks (CNNs), shows promise in accurately classifying breast cancer images. However, CNNs face challenges with hyperparameters. To address these challenges, this article proposes a new DL model, the optimized DenseNet model (LFR-COA-DenseNet121-BC), incorporating a boosted metaheuristic algorithm called LFR-COA. This algorithm, a developed version of the Coati Optimization Algorithm (COA), integrates Random opposition-based learning (ROB), Brownian motion, and Lévy Flight (LF) schemes. The proposed model achieved impressive results, accurately classifying 99.97% of the test set. Comparison with established models such as VGG16, VGG19, DenseNet201, InceptionV3, Xception, and MobileNet revealed superior performance of the LFR-COA-DenseNet121-BC model, with 99.97% accuracy, 99.96% sensitivity, and 99.9% specificity. Further comparison with COA-DenseNet121 highlighted the superiority of LFR-COA-DenseNet-BC. In addition, LFR-COA was evaluated at the IEEE Congress on Evolutionary Computation held in 2022 (CEC 2022), and real-world medical scenarios showcased the effectiveness of LFR-COA compared to existing optimization algorithms. LFR-COA consistently outperformed the original COA algorithm and other well-known counterparts in various statistical, convergence, and diversity measures, affirming its efficacy in breast cancer classification.

Detectability of Breast Cancer Through Inverse Heat Transfer Modeling Using Patient-Specific Surface Temperatures

Abstract Breast cancer in women is a prevalent disease which takes over 680,000 lives each year worldwide. Early detection of breast cancer through screening has played a significant role in reducing the mortality rates. The current screening paradigm has shown the difficulties in detecting cancers for patients with dense breasts, small and deep tumors, and cancer types that are difficult to visualize. Infrared imaging (IRI) aided by advanced thermal analysis of the breast has shown great promise in detecting cancer using surface temperatures effected by a metabolically active and highly perfused tumor region. We previously developed an inverse heat transfer approach to detect the presence and absence of breast cancer using IRI, called the IRI-Numerical Engine. It was validated with 23 biopsy-proven breast cancer patients irrespective of breast density and cancer type at various tumor depths (0.95 cm to 5.45 cm from the breast surface). The current work is aimed to obtain the detectability limit of the IRI-Numerical Engine by testing the capability of detecting 10-20 mm tumors at various depths in patient-specific digital breast models. In addition, a study on the effect of tumor size, tumor location, breast shape, and breast size on the surface temperatures of patient-specific models was conducted to verify that an IR camera could capture these surface temperature distributions. The algorithm was able to detect the presence of a tumor at various depths, and deep tumors are detectable given the appropriate thermal sensitive IR camera.

DCFNet: Infrared and Visible Image Fusion Network Based on Discrete Wavelet Transform and Convolutional Neural Network.

Aiming to address the issues of missing detailed information, the blurring of significant target information, and poor visual effects in current image fusion algorithms, this paper proposes an infrared and visible-light image fusion algorithm based on discrete wavelet transform and convolutional neural networks. Our backbone network is an autoencoder. A DWT layer is embedded in the encoder to optimize frequency-domain feature extraction and prevent information loss, and a bottleneck residual block and a coordinate attention mechanism are introduced to enhance the ability to capture and characterize the low- and high-frequency feature information; an IDWT layer is embedded in the decoder to achieve the feature reconstruction of the fused frequencies; the fusion strategy adopts the l1-norm fusion strategy to integrate the encoder's output frequency mapping features; a weighted loss containing pixel loss, gradient loss, and structural loss is constructed for optimizing network training. DWT decomposes the image into sub-bands at different scales, including low-frequency sub-bands and high-frequency sub-bands. The low-frequency sub-bands contain the structural information of the image, which corresponds to the important target information, while the high-frequency sub-bands contain the detail information, such as edge and texture information. Through IDWT, the low-frequency sub-bands that contain important target information are synthesized with the high-frequency sub-bands that enhance the details, ensuring that the important target information and texture details are clearly visible in the reconstructed image. The whole process is able to reconstruct the information of different frequency sub-bands back into the image non-destructively, so that the fused image appears natural and harmonious visually. Experimental results on public datasets show that the fusion algorithm performs well according to both subjective and objective evaluation criteria and that the fused image is clearer and contains more scene information, which verifies the effectiveness of the algorithm, and the results of the generalization experiments also show that our network has good generalization ability.

Investigation of spectral bands and sensor parameters for methane emission detection imaging spectrometer

Recently, interest has increased in developing remote sensing (RS) sensor systems with a high spatial and spectral resolution for quantifying methane (CH4) emissions from point sources. Evaluating sensor parameters can lead to a better understanding of technological advancements in CH4 emission estimation. This study addresses several important topics, including the robust algorithm for point source CH4 emission detection, selecting optimum CH4 absorption bands and their sensitivity analysis, and evaluating sensor parameters such as spatial and spectral resolution and signal-to-noise ratio (SNR). Since the matched filter algorithm efficiently detects point source emissions with palpable accuracy, it has been used here for the Airborne Visible InfraRed Imaging Spectrometer – Next Generation (AVIRIS-NG) sensor data acquired over the USA. The optimum band selection for CH4 detection was determined using a normalized sensitivity method to investigate the effects of aerosol, water vapor (WV), and surface albedo variations. The impact of these parameters on the CH4 absorption bands in the short-wave infrared (SWIR) range has been analyzed using MODerate resolution atmospheric TRANsmission (MODTRAN) radiative transfer (RT) model simulations. Using AVIRIS-NG data as a testbed, various methods have been used to reconstruct original data and generate data sets with different sensor specifications. The effects of the sensor parameters, such as spatial and spectral resolution and SNR, on the detection accuracy were analyzed. The results show the significance of SNR over the spatial and spectral resolution for better emission detection. It indicates the reasonable selection of sensor parameters that can help to create an optimal sensor system that will ultimately support the development of an imaging spectrometer suitable for CH4 emission estimation.

High-resolution thermal infrared contrails images identification and classification method based on SDGSAT-1

With the rapid development of the aviation industry, the impact of contrails on global warming is becoming increasingly significant. However, due to limitations in sensor technology and data sources, precise, all-weather observations of contrails remain a significant challenge. To address this, this study utilizes thermal imagery data from the SDGSAT-1 to classify newly formed contrails based on their length, creating SDGCONTR, the first three-band thermal infrared contrail dataset with a 30-meter spatial resolution. Based on this dataset, we propose an effective contrail image classification method: the SDGContrail Images Classification Method. This method builds on the existing Multi-axis Vision Transformer deep learning model. In the preprocessing stage, it employs an innovative contrail enhancement method by creating masks based on contrail brightness temperature and brightness temperature ratio percentiles in thermal infrared images. Our method achieved optimal classification performance compared to other methods, with the F1 Score of 95.2%. Using this model, we further analyzed the geographic and diurnal distribution patterns of long and short contrails, validating the mechanism of contrail formation. Our method not only enhances the ability to conduct large-scale, all-weather, and detailed contrail observations, but also sets a new standard for environmental monitoring and aviation safety.

Electroluminescence and infrared imaging of fielded photovoltaic modules: A complementary analysis of series resistance-related defects

Fielded photovoltaic (PV) modules often exhibit series resistance-related defects, mainly due to solder bond degradation and metallization corrosion. To quantitatively analyze the series resistance increase in fielded modules, we conducted a detailed investigation on two sets of modules, employing two complementary techniques, electroluminescence (EL) and infrared (IR) imaging. The dependence of EL image characteristics on module temperature, as revealed in IR images, was a key focus of this study. The first and second sets of modules were exposed in Florida (hot and humid climate) and Arizona (hot and dry climate) over 10 years and 18 years, respectively. The resistive defect patterns obtained using EL and IR images showed a closer correlation for the Florida modules compared to the Arizona modules as the Florida modules primarily experience solder bond degradation. EL and IR images were acquired at five current injection levels (i.e., 0.1, 0.3, 0.5, 0.7, 1.0 x short circuit current) and two exposure times (i.e., 60 s and 300 s) and used to develop and report a new curve fitting method for estimating the external series resistance. The results indicate that inaccurate temperature determinations from IR images can lead to underestimations (up to 23 %) in EL-based external series resistance estimates. For the most accurate series resistance estimation, especially in modules with severe thermal defects and series resistance deterioration, the study recommends obtaining EL and IR images within 60 s of the current injection time. This study also reports a Monte Carlo simulation assessing the impact of EL and IR characteristics on the accuracy of external series resistance estimations.

A calorimetric evaluation method for beam targets with IR imaging: Implementation for the negative ion source BATMAN Upgrade

A comprehensive calorimetric evaluation method is presented for the beam target installed in the testbed BATMAN Upgrade (BUG). BUG hosts the prototype RF-driven negative ion source for the ITER Neutral Beam Heating System. In nominal conditions BUG features an array of negative ion beamlets (5 horizontal × 14 vertical) with 20 mm spacing among them. Each beamlet is accelerated electrostatically up to a maximum of 45 kV and transports currents in the range from 0.01 to 0.05 A. BUG has two beam targets at different distances from the beam acceleration system. The furthest (2.08 m downstream) is a water-cooled diagnostic CW-calorimeter, which can measure the heat flux distribution of the beam with a resolution of 20 × 20 mm² over a surface of 600 (H) × 800 (V) mm² and water calorimetry. The closest (0.86 m downstream) beam target with sub-millimetric resolution consists of a (376 × 142 × 20 mm³) tile made of a composite anisotropic material with carbon fibers oriented in the thickness direction (1D-CFC). It can only be exposed to beams for limited time, therefore it is retractable. Infrared (IR) imaging of beam targets yields time-resolved high-resolution spatial temperature footprints. A simple algorithm is proposed to reconstruct the heat flux at the front of a beam target, from the IR footprints at the rear. The reconstructed heat flux profiles are lower and broader due to heat conduction effects in the beam target. A correction algorithm is proposed through simulation-based inference of 2D axisymmetrical FEM simulations that reduces the power and shape fit errors over an order of magnitude, to a few percent. The calorimetric evaluation method is applied to experimental shots and two advanced heat flux distribution analysis are shown. The most relevant error sources are discussed and quantified to assess systematic errors. Finally, the method is validated in BUG comparing total beam power estimations with the two calorimetric diagnostics from the CW-calorimeter, showing good agreement.

Visualization of Keratopathy Associated With the Antibody-Drug Conjugate Belantamab Mafodotin Using Infrared Imaging in Patients With Multiple Myeloma.

The treatment of patients with relapsed/refractory multiple myeloma (RRMM) with the antibody-drug conjugate belantamab mafodotin is affected by ocular adverse effects, most frequently keratopathy with corneal microcyst-like epithelial changes (MECs). To assess ocular side effects, the "Keratopathy and Visual Acuity (KVA) scale," based on the extent of keratopathy subjectively graded on slit-lamp examination and the change in best corrected visual acuity from baseline, was created. Advanced corneal imaging techniques have been explored to further characterize MECs and identify objective imaging biomarkers. We examined whether infrared reflectance imaging of the anterior segment (AS-IR) could contribute to the assessment, monitoring, and documentation of corneal toxicity in patients treated with belantamab mafodotin. In addition to the KVA examination, AS-IR imaging was performed. AS-IR images were evaluated for presence of visible hyporeflective lesions and their spatial and temporal distribution between visits and compared with keratopathy identified on slit-lamp examination. To standardize the assessment, a scoring system for lesions on AS-IR was implemented for additional analysis. Nine patients undergoing treatment with belantamab mafodotin for up to 9 months were examined. All patients exhibited hyporeflective lesions on AS-IR imaging, indicative of corneal toxicity corresponding to MECs observed on slit-lamp examination. AS-IR lesions showed early occurrence, variable quantity and size, and distinct distribution patterns, correlating with clinical findings during treatment. As shown for belantamab mafodotin, AS-IR imaging represents a fast, noninvasive, supplemental method for documentation, monitoring, and assessment of corneal adverse effects during treatment with antibody-drug conjugates, which may enable more standardized analyses.

BeO in epoxy composites as TIM for high-power LED packages

Beryllium oxide (BeO)-filled epoxy composite was prepared and used as a thermal interface material (TIM) in light-emitting diode (LED) packages. The filler percentages influenced on changing thermal and optical behaviour of LED at various driving currents. An improvement of 14.36% in thermal conductivity was recorded from that of neat epoxy. The lowest total thermal resistance of 10.75 K/W ( Rth) and rise in junction temperature ( Tj) of 3.1°C were recorded at 100 mA driving current for 50 wt-% BeO-filled TIM attached LED followed by an improved optical output and lowered correlated colour temperature values (3466 K). Thermal imaging infra-red imaging analysis also supported the observation of reduced Tj and Rth values from transient analysis and indicated lower chip surface temperature ( Ts) at all driving currents.

Design of A prototype 128 × 128 ROIC array for 2.6 μm-wavelength SWIR image sensor applications

This paper presents the development and evaluation of a 128 × 128 Readout Integrated Circuit (ROIC) prototype, engineered for Short-Wave Infrared (SWIR) imaging at a specific target wavelength of 2.6 μm. Employing silicon-level verification, this work undertook an exhaustive analysis of the ROIC's performance, identifying key areas for enhancement to improve SWIR imaging systems. Fabricated with 0.18-μm CMOS technology, the ROIC is tailored for integration with Indium Gallium Arsenide (InGaAs) Focal Plane Arrays (FPAs), facilitating high-resolution imaging. The prototype consumes 42.25 mW of power and achieves a frame rate of 390 frames per second. The fabricated chip show that the random noise level is 72.65 μVrms and Pixel-FPN is 21 LSBrms. This investigation lays a critical groundwork for future SWIR imaging advancements, providing valuable insights and methodologies to boost imaging performance in various applications.

AI-Based Pedestrian Detection and Avoidance at Night Using Multiple Sensors

In this paper, we present a pedestrian detection and avoidance scheme utilizing multi-sensor data collection and machine learning for intelligent transportation systems (ITSs). The system integrates a video camera, an infrared (IR) camera, and a micro-Doppler radar for data acquisition and training. A deep convolutional neural network (DCNN) is employed to process RGB and IR images. The RGB dataset comprises 1200 images (600 with pedestrians and 600 without), while the IR dataset includes 1000 images (500 with pedestrians and 500 without), 85% of which were captured at night. Two distinct DCNNs were trained using these datasets, achieving a validation accuracy of 99.6% with the RGB camera and 97.3% with the IR camera. The radar sensor determines the pedestrian’s range and direction of travel. Experimental evaluations conducted in a vehicle demonstrated that the multi-sensor detection scheme effectively triggers a warning signal to a vibrating motor on the steering wheel and displays a warning message on the passenger’s touchscreen computer when a pedestrian is detected in potential danger. This system operates efficiently both during the day and at night.

Synergistic retrieval of mangrove vital functional traits using field hyperspectral and satellite data

Mangroves play a pivotal role in the attainment of United Nations Sustainable Development Goals. Leaf equivalent water thickness (EWT) and leaf mass per area (LMA) are vital functional trait parameters for monitoring heath status of mangroves. However, their inherent spectral response characteristics remain unclear, which presents a challenge for high-precision estimating EWT and LMA using remote sensing images. In this study, we attempted to synergistic retrieval of mangrove EWT and LMA using field hyperspectral and satellite data. Specifically, we proposed a novel Fractional-order assisted Spectral angle Analysis (FSA) approach to accurately identify sensitive spectral domains of EWT and LMA for three mangrove species based on in situ full-spectrum hyperspectral (350–2500 nm) data. We constructed a new spectral matching index (SMI) to screen the satellite-based sensitive spectral bands based on the field-based spectral domains, and further verified their feasibility of estimating EWT and LMA using three optical satellite sensors (Sentinel-2A, SDGSAT-1, and OHS-3D). Finally, we explored the effectiveness of combining SAR and thermal infrared (TIR) images with sensitive spectral bands on improving retrieval performance in estimating EWT and LMA, respectively. Results indicated that FSA approach could capture the sensitive spectral domains of EWT (1125–1868 nm) and LMA (1419–1999 nm) for three mangrove species. The SMI could effectively realize the spectral matching between field-based sensitive domains and satellite-based bands, and perform the synergistic retrieval of mangrove EWT (R2 = 0.65) and LMA (R2 = 0.70). Intriguing finding is that the additional TIR bands significantly enhanced the EWT and LMA inversion accuracy (R2 increased by 9.8 %–68.3 %) when compared to SAR images. We confirmed that EWT displayed an inverse trend of LMA with the flooding frequency increasing. Our works provide a scientific basis for protection and sustainable management of mangroves.

Diagnosing uncertainties in global biomass burning emission inventories and their impact on modeled air pollutants

Abstract. Large uncertainties persist within current biomass burning (BB) inventories, and the choice of these inventories can substantially impact model results when assessing the influence of BB aerosols on weather and climate. We evaluated discrepancies among BB emission inventories by comparing carbon monoxide (CO) and organic carbon (OC) emissions from seven major BB regions globally between 2013 and 2016. Mainstream bottom-up inventories, including the Fire INventory from NCAR 1.5 (FINN1.5) and Global Fire Emissions Database version 4s (GFED4s), along with the top-down inventories Quick Fire Emissions Dataset 2.5 (QFED2.5) and Visible Infrared Imaging Radiometer (VIIRS-)based Fire Emission Inventory version 0 (VFEI0), were selected for this study. Global CO emissions range from 252 to 336 Tg, with regional disparities reaching up to a 6-fold difference. Dry matter is the primary contributor to the regional variation in CO emissions (50 %–80 %), with emission factors accounting for the remaining 20 %–50 %. Uncertainties in dry matter often arise from biases in calculating bottom fuel consumption and burned area, influenced by vegetation classification methods and fire detection products. In the tropics, peatlands contribute more fuel loads and higher emission factors than grasslands. At high latitudes, increased cloud fraction amplifies the discrepancy in estimated burned area (or fire radiative power) by 20 %. The global OC emissions range from 14.9 to 42.9 Tg, exhibiting higher variability than CO emissions due to the corrected emission factors in QFED2.5, with regional disparities reaching a factor of 8.7. Additionally, we applied these BB emission inventories to the Community Atmosphere Model version 6 (CAM6) and assessed the model performance against observations. Our results suggest that the simulations based on the GFED4s agree best with the Measurements of Pollution in the Troposphere (MOPITT-)retrieved CO. While comparing the simulation with the Moderate Resolution Imaging Spectroradiometer (MODIS) and AErosol RObotic NETwork (AERONET) aerosol optical depth (AOD), our results reveal that there is no global optimal choice for BB inventories. In the high latitudes of the Northern Hemisphere, using GFED4s and QFED2.5 can better capture the AOD magnitude and diurnal variation. In equatorial Asia, GFED4s outperforms other models in representing day-to-day changes, particularly during intense burning. In Southeast Asia, we recommend using the OC emission magnitude from FINN1.5 combined with daily variability from QFED2.5. In the Southern Hemisphere, the latest VFEI0 has performed relatively well. This study has implications for reducing the uncertainties in emissions and improving BB emission inventories in further studies.

Pilot-scale performance of gravity-driven ultra-high flux fabric membrane systems for removing small-sized microplastics in wastewater treatment plant effluents

The ubiquitous nature and environmental impacts of microplastic particles and fibers demand effective solutions to remove such micropollutants from sizable point sources, including wastewater treatment plants and road runoff facilities. While advanced methods, e.g., microfiltration and ultrafiltration, have shown high removal efficiencies of small-sized microplastics (<150 μm), the low flux encountered in these systems implies high operation costs and makes them less effective in high-capacity wastewater facilities. The issue presents new opportunities for developing cheap high-flux membrane systems, deployable in low-to high-income economies, to remove small-sized microplastic and nanoplastics in wastewater. Here, we report on developing an ultra-high flux gravity-driven fabric membrane system, assessed through a laboratory-scale filtration and large-scale performance in an actual wastewater treatment plant (WWTP). The method followed a carefully designed water sampling, pre-treatment protocol, and analytical measurements involving Fourier transform infrared (FTIR) spectroscopy and laser direct infrared (LDIR) imaging. The result shows that the ultra-high flux (permeance = 550,000 L/m2h⋅bar) fabric membrane system can effectively remove small-sized microplastics (10–300 μm) in the secondary effluent of an actual WWTP at high efficiency greater than 96 %. The pilot system demonstrated a continuous treatment capacity of 300,000 L/day through a 1 m2 surface area disc, with steady removal rates of microplastics. These findings demonstrate the practical, cheap, and sustainable removal of small-sized microplastics in wastewater treatment plants, and their potential value for other large-scale point sources, e.g., stormwater treatment facilities.

Using Passive Multi-Modal Sensor Data for Thermal Simulation of Urban Surfaces

Abstract. This paper showcases an integrated workflow hinged on passive airborne multi-modal sensor data for the simulation of the thermal behavior of built-up areas with a focus on urban heat islands. The geometry of the underlying parametrized model, or digital twin, is derived from high-resolution nadir and oblique RGB, near-infrared and thermal infrared imagery. The captured bitmaps get photogrammetrically processed into comprehensive surface models, terrain, dense 3D point clouds and true-ortho mosaics. Building geometries are reconstructed from the projected point sets with procedures presupposing outlining, analysis of roof and fac¸ade details, triangulation, and texturing mapping. For thermal simulation, the composition of the ground is determined using supervised machine learning based on a modified multi-modal DeepLab v3+ architecture. Vegetation is retrieved as individual trees and larger tree regions to be added to the meshed terrain. Building materials are assigned from the available visual, infrared and surface planarity information as well as publicly available references. With actual weather data, surface temperatures can be calculated for any period of time by evaluating conductive, convective, radiative and emissive energy fluxes for triangular layers congruent to the faces of the modeled scene. Results on a sample dataset of the Moabit district in Berlin, Germany, showed the ability of the simulator to output surface temperatures of relatively large datasets efficiently. Compared to the thermal infrared images, several insufficiencies in terms of data and model caused occasional deviations between measured and simulated temperatures. For some of these shortcomings, improvement suggestions within future work are presented.

DeDNet: Infrared and visible image fusion with noise removal by decomposition-driven network

Infrared and visible image fusion integrates useful information from different modal images to generate one with comprehensive details and highlighted targets, thereby deepening the interpretation of the scene. However, existing deep-learning based methods do not consider noise, leading to suboptimal noisy fusion results. To address this issue, we propose a decomposition-driven neural network (DeDNet) to achieve joint fusion and noise removal. By introducing constraints between the fused and ground truth source images into the loss function, we develop an autoencoder as the basic fusion and denoising network. Furthermore, we propose a decomposition network that guided the decomposition of the fusion result, improving the denoising and details recovery. Experiments demonstrate DeDNet excels the state-of-the-art methods in objective and subjective evaluations, yielding competing performance in detection and segmentation. On the metrics,Qcb, EN, SSIM, PSNR, and CC, DeDNet average increased 10.92%, 21.13%, 82.97%, 8.55%, and 16.26% than the compared methods, respectively. The source code is available at https://github.com/JasonWong30/DeDNet.

DustSCAN: A Five Year (2018-2022) Hourly Dataset of Dust Plumes From SEVIRI

Airborne mineral dust significantly impacts air quality, human health, and the global climate. Due to sparse ground sensors, particularly in source regions, dust monitoring relies mainly on remote sensing through Aerosol Optical Depth (AOD) retrievals from polar-orbiting satellite optical instruments. These are valuable but lack the temporal resolution for precise plume tracking and source characterization. We introduce DustSCAN, a five-year, hourly dust plume dataset derived from the Spinning Enhanced Visible and InfraRed Imager (SEVIRI) images on geostationary-orbit Meteosat satellites. Using multi-channel infrared images, we detect atmospheric dust and track hourly dust-affected pixels. These are clustered into discrete plumes using the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm. DustSCAN includes 9950 discrete plumes over 2018-2022 across the Sahara, the Arabian Desert, and Western and Central Asia. It complements existing resources and provides a framework for detailed analysis of dust sources, trajectories, and impacts. Its distinctive event-based and spatio-temporal detail offers an advancement in unraveling the complexities of dust storm dynamics.

Thermal Infrared-Image-Enhancement Algorithm Based on Multi-Scale Guided Filtering

Obtaining thermal infrared images with prominent details, high contrast, and minimal background noise has always been a focal point of infrared technology research. To address issues such as the blurriness of details and low contrast in thermal infrared images, an enhancement algorithm for thermal infrared images based on multi-scale guided filtering is proposed. This algorithm fully leverages the excellent edge-preserving characteristics of guided filtering and the multi-scale nature of the edge details in thermal infrared images. It uses multi-scale guided filtering to decompose each thermal infrared image into multiple scales of detail layers and a base layer. Then, CLAHE is employed to compress the grayscale and enhance the contrast of the base layer image. Then, detail-enhancement processing of the multi-scale detail layers is performed. Finally, the base layer and the multi-scale detail layers are linearly fused to obtain an enhanced thermal infrared image. Our experimental results indicate that, compared to other methods, the proposed method can effectively enhance image contrast and enrich image details, and has higher image quality and stronger scene adaptability.

Wall heating by subcritical energetic electrons generated by the runaway electron avalanche source *

Subcritical energetic electrons (SEEs) produced by the runaway electron (RE) avalanche source at energies below the runaway threshold are found to be the primary contributor to surface heating of plasma-facing components (PFCs) during final loss events. This finding is supported by theoretical analysis, computational modeling with the Kinetic Orbit Runaway electrons Code (KORC), and qualitative agreement with DIII-D experimental observations. The avalanche source generates significantly more secondary electrons below the runaway threshold, which thermalize rapidly when well-confined. However, during a final loss event, the RE beam impacts the first wall, and SEEs are deconfined before they can thermalize. Additionally, because the energy deposition length decreases faster than energy, the deposited energy density, and thus the maximum PFC surface temperature change, is larger for SEEs than REs. KORC simulations employ an analytic first wall to model particle deconfinement onto a non-axisymmetric wall composed of individual tiles. PFC surface heating is calculated using a 1D model extended to include an energy-dependent deposition length scale. Simulations of DIII-D qualitatively agree with infrared (IR) imaging only when SEEs from the avalanche source are included. These results demonstrate that SEEs are the dominant contributor to PFC surface heating and indicate that the avalanche source plays a critical role in the PFC damage caused during final loss events. The prominence of SEEs also has important implications for interpreting IR imaging, one of the primary diagnostics for RE-wall interaction diagnosis, despite REs dominating the energy and current density. This result improves predictions of wall damage due to post-disruption REs to estimate material lifetime and design RE mitigation systems for ITER and future reactors.