Thermal Anomaly Detection Research Articles

Vision transformer-enhanced thermal anomaly detection in building facades through fusion of thermal and visible imagery

ABSTRACT The thermal anomaly detection of building facades is critically important for the evaluation and upkeep of structures. While conventional methods are effective in numerous instances, their capability to discern flaws in complex environments is constrained. This research employs an optimized Vision Transformer (ViT) technology, enhanced by the Line Acquisition, Filtering, and Revision (LAFR) algorithm. By amalgamating the distinctive advantages of ViT’s attention-based mechanisms with the complementary data from thermal and visible imagery, this approach significantly bolsters the detection of thermal anomalies on building facades. The experimental outcomes reveal that our methodology, with an F1 score of 0.847, markedly surpasses traditional thermal imaging (F1 score of 0.466), showcasing its potential for precise anomaly detection in architectural contexts. This validates the method as a more robust and efficient solution for structural evaluation, promising to elevate the quality of architectural inspections and maintenance.

Satellite Time-Series Analysis for Thermal Anomaly Detection in the Naples Urban Area, Italy

Naples is the most densely populated Italian city (7744 inhabitants per km2). It is located in a particular geological context: the presence of Mt Vesuvius characterizes the eastern part, and the western part is characterized by the presence of the Phlegrean Fields, making Naples a high-geothermal-gradient region. This endogenous heat, combined with the anthropogenic heat due to intense urbanization, has defined Naples as an ideal location for Surface Urban Heat Island (SUHI) analysis. SUHI analysis was effectuated by acquiring the Land Surface Temperature (LST) over Naples municipality by processing Landsat 8 (L8) Thermal Infrared Sensor (TIRS) images in the 2013–2023 time series by employing Google Earth Engine (GEE). In GEE, two different approaches have been followed to analyze thermal images, starting from the Statistical Mono Window (SMW) algorithm, which computes the LST based on the brightness temperature (Tb), the emissivity value, and the atmospheric correction coefficients. The first one is used for the LST retrieval from daytime images; here, the emissivity component is derived using, firstly, the Normalized Difference Vegetation Index (NDVI) and then the Vegetation Cover Method (VCM), defining the Land Surface Emissivity (LSɛ), which considers solar radiation as the main source of energy. The second approach is used for the LST retrieval from nighttime images, where the emissivity is directly estimated from the Advance Spaceborne Thermal Emission Radiometer database (ASTER-GED), as, during nighttime without solar radiation, the main source of energy is the energy emitted by the Earth’s surface. From these two different algorithms, 123 usable daytime and nighttime LST images were downloaded from GEE and analyzed in Quantum GIS (QGIS). The results show that the SUHI is more concentrated in the eastern part, characterized by intense urbanization, as shown by the Corine Land Cover (CLC). At the same time, lower SUHI intensity is detected in the western part, defined by the Land Cover (LC) vegetated class. Also, in the analysis, we highlighted 40 spots (10 hotspots and 10 coldspots, both for daytime and nighttime collection) that present positive or negative temperature peaks for all the time series. Due to the huge amount of data, this work considered only the five representative spots that were most representative for SUHI analysis and determination of thermal anomalies in the urban environment.

Study on the Preparation and Aging Performance of Temperature-Indicating Patch Used for Thermal Defect Detection of Transformer Bushing Cylinder Head

Elevated temperatures at the transformer bushing cylinder head can precipitate failures, leading to significant power outages. In response, this study introduces a reversible temperature-indicating patch for the nuanced detection of thermal anomalies in the transformer bushing’s cylinder head. The patch, crafted through a melting process, utilizes a reversible discoloration material and is developed in two variants via an adsorption substrate method. Comprehensive evaluations of the patches’ color-changing characteristics, alongside their electrical and hydrophobic properties, were conducted using an automatic contact angle measuring instrument and an AC flashover test platform. The findings reveal that the temperature-indicating patch exhibits a discernible color transition within the range of 49~55 °C, with a color reversion temperature span of 45~55 °C, denoting marked sensitivity and robust reversibility. Additionally, it was observed that prolonged thermal aging correlates with a decrease in both the water contact angle and the discharge voltage per unit length across the surface of the patches, indicating a degradation in performance. Among the variants, the binder-based temperature-indicating patch demonstrated superior stability in electrical performance compared to its vacuum-based counterpart. The outcomes of this research offer valuable insights for the development of advanced diagnostic tools for the identification of thermal defects in transformer bushings, potentially enhancing reliability and safety in power distribution systems.

Low-Cost Distributed Thermal Response Test for the Estimation of Thermal Ground and Grout Conductivities in Geothermal Heat Pump Applications

The design process of a borehole heat exchanger (BHE) requires knowledge of building thermal loads, the expected heat pump’s COP and the ground’s thermophysical properties. The thermal response test (TRT) is a common experimental technique for estimating the ground’s thermal conductivity and borehole thermal resistance. In classic TRT, a constant heat transfer rate is provided above ground to the carrier fluid that circulates continuously inside a pilot BHE. The average fluid temperature is measured, and from its time-dependent evolution, it is possible to infer both the thermal resistance of the BHE and the thermal conductivity of the ground. The present paper investigates the possibility of a new approach for TRT with the continuous injection of heat directly into the BHE’s grouting by means of electrical resistance imparted along the entire BHE’s length, while local (along the depth) temperature measurements are acquired. This DTRT (distributed TRT) approach has seldom been applied and, in most applications, circulating hot fluid and optical fibers are used to infer depth-related temperatures. The distributed measurements allow the detection of thermal ground anomalies along the heat exchanger and even the presence of aquifer layers. The present paper investigates the new EDDTRT (electric depth-distributed TRT, under patenting) approach based on traditional instruments (e.g., RTD) or one-wire digital sensors. The accuracy of the proposed method is numerically assessed by Comsol Multiphysics simulations. The analysis of the data obtained from the “virtual” EDDTRT confirms the possibility of estimating within 10% accuracy both thermal ground and grout conductivities.

LONG-TERM THERMAL ANOMALY DETECTION AND MAPPING AT PIXEL LEVEL USING A GOOGLE EARTH ENGINE TOOL

Abstract. Frequency of extreme weather events such as cloudbursts, heatwaves etc. have increased as an outcome of changing climate. Identification of the pattern of extreme temperature events is important since it governs various events such as heatwaves, wildfires, droughts, storms, coldwaves etc. Moderate Resolution Imaging Spectroradiometer (MODIS) provides Land Surface Temperature (LST) data at 1 kilometre of spatial resolution at daily interval that can help in the identification and mapping of the anomalies in the temperature at pixel level. This study proposes a global-scale daily long-term thermal anomaly detection tool made using Google Earth Engine (GEE) App. This open source tool with the name of ‘Deviation from Mean’ uses the MODIS LST data available from 2000 till date to detect temperature anomaly based on the deviation of temperature of any day (chosen by the user) from the long-term climatological mean. It also generates a time-series plot of temperature values of any pixel for any date for last 24 years i.e. 2000–2023 in the graphical form to analyze the variation in the temperature over the time. A case study has also been done using the tool to highlight the thermal anomaly experienced over the Indian sub-continent during March-April, 2022 and 2023. This tool is capable of providing thermal anomaly information at global, regional as well as local level that can help in taking region-specific mitigation measures.

AI-powered inspections of facades in reinforced concrete buildings

Worldwide there are plenty of aged Reinforced Concrete (RC) buildings in need of thorough inspections. Cracks, delamination, stains, leakages, debonding and moisture ingressions are common defects found in RC structures. Such problems are typically diagnosed through qualitative assessment of visual and thermal photographs (data) by certified inspectors. However, qualitative inspections are very tedious, time-consuming and costly. This paper presents an alternative novel approach to drastically increase efficiency by decreasing the data collection and analysis time. Data collection for the inspection of facades is undertaken with Unmanned Aerial Vehicles (UAVs) either through an autonomous pre-programmed flight or through a human-piloted flight. Data analysis is performed by implementing up-to-date AI-powered algorithms to automatically detect defects on visual and thermal photographs. All the recognised defects and thermal anomalies are labelled on the building facade for comprehensive evaluation of the asset. This paper reports that the implementation of AIpowered inspections can save up to 67% of the time spent and 52% of the cost in comparison to the most commonly adopted practice in the industry with an average accuracy of 90.5% and 82% for detection of visual defects and thermal anomalies, respectively.

Data-Driven Thermal Anomaly Detection in Large Battery Packs

The early detection and tracing of anomalous operations in battery packs are critical to improving performance and ensuring safety. This paper presents a data-driven approach for online anomaly detection in battery packs that uses real-time voltage and temperature data from multiple Li-ion battery cells. Mean-based residuals are generated for cell groups and evaluated using Principal Component Analysis. The evaluated residuals are then thresholded using a cumulative sum control chart to detect anomalies. The mild external short circuits associated with cell balancing are detected in the voltage signals and necessitate voltage retraining after balancing. Temperature residuals prove to be critical, enabling anomaly detection of module balancing events within 14 min that are unobservable from the voltage residuals. Statistical testing of the proposed approach is performed on the experimental data from a battery electric locomotive injected with model-based anomalies. The proposed anomaly detection approach has a low false-positive rate and accurately detects and traces the synthetic voltage and temperature anomalies. The performance of the proposed approach compared with direct thresholding of mean-based residuals shows a 56% faster detection time, 42% fewer false negatives, and 60% fewer missed anomalies while maintaining a comparable false-positive rate.

Detection and Identification of Preseismic Thermal Anomalies in Cloudy Conditions Associated With the 2022 Luding (China) Mw 6.6 Earthquake

Detection and Identification of Preseismic Thermal Anomalies in Cloudy Conditions Associated With the 2022 Luding (China) Mw 6.6 Earthquake

A spatio-temporal temperature-based thresholding algorithm for underground coal fire detection with satellite thermal infrared and radar remote sensing

Coal fire is a global catastrophe that causes resource waste, economic loss, and environmental pollution. Accurate detection of surface abnormalities induced by coal fires (such as thermal anomalies and land deformation) is critical for coal fire monitoring and extinguishment. However, the conventional thermal anomaly detection methods do not consider the temporal variation of land surface temperature (LST). Moreover, previous studies have not applied the polarimetric persistent scatterer interferometry(PolPSI) technique for coal-fire-related ground deformation monitoring, which is supposed to have better performance than conventional persistent scatterer interferometry (PSI) approaches. To this end, in this study, a spatio-temporal temperature-based thresholding (STTBT) algorithm is proposed to detect thermal anomalies, firstly. Then, the ground deformation obtained by the PolPSI method is combined with the detected thermal anomalies for the coal fire detection. Besides, a two-stage bandpass filter is employed to assist in detecting the suspected coal fire. The results in the Fukang coal fire area, China, demonstrate that, compared with the conventional methods, the proposed STTBT mothed achieves higher accuracy of thermal anomalies detection, thereby can able to detect coal fire areas more effectively. In addition, PolPSI has better performance than PSI in ground deformation monitoring over coal fire areas. The proposed STTBT combined with PolPSI can serve as a powerful tool for detecting and monitoring underground coal fires.

The Impact of Dynamic Emissivity–Temperature Trends on Spaceborne Data: Applications to the 2001 Mount Etna Eruption

Spaceborne detection and measurements of high-temperature thermal anomalies enable monitoring and forecasts of lava flow propagation. The accuracy of such thermal estimates relies on the knowledge of input parameters, such as emissivity, which notably affects computation of temperature, radiant heat flux, and subsequent analyses (e.g., effusion rate and lava flow distance to run) that rely on the accuracy of observations. To address the deficit of field and laboratory-based emissivity data for inverse and forward modelling, we measured the emissivity of ‘a’a lava samples from the 2001 Mt. Etna eruption, over the wide range of temperatures (773 to 1373 K) and wavelengths (2.17 to 21.0 µm). The results show that emissivity is not only wavelength dependent, but it also increases non-linearly with cooling, revealing considerably lower values than those typically assumed for basalts. This new evidence showed the largest and smallest increase in average emissivity during cooling in the MIR and TIR regions (~30% and ~8% respectively), whereas the shorter wavelengths of the SWIR region showed a moderate increase (~15%). These results applied to spaceborne data confirm that the variable emissivity-derived radiant heat flux is greater than the constant emissivity assumption. For the differences between the radiant heat flux in the case of variable and constant emissivity, we found the median value is 0.06, whereas the 25th and the 75th percentiles are 0.014 and 0.161, respectively. This new evidence has significant impacts on the modelling of lava flow simulations, causing a dissimilarity between the two emissivity approaches of ~16% in the final area and ~7% in the maximum thickness. The multicomponent emissivity input provides means for ‘best practice’ scenario when accurate data required. The novel approach developed here can be used to test an improved version of existing multi-platform, multi-payload volcano monitoring systems.

Re2FAD: A differential image registration and robust image fusion method framework for power thermal anomaly detection

Previous image fusion methods may not effectively characterize thermal anomalies in power equipment fault detection. In addition, the source images of power equipment collected on-site are not registered and cannot be directly used for image fusion. To these problems, this paper proposes a framework for differential image registration and robust image fusion to assist thermal anomaly detection (Re2FAD). First, Re2FAD calculates a joint histogram of the source images after low-frequency decomposition and constructs a criterion function to solve the optimal registration parameters. Second, the membership function is constructed to clarify the labels of the segmented regions. Third, the fusion strategy is selected according to the label to reconstruct the fused image. We built the power equipment dataset of the Shanghai Sanlin 500kV substation and proved that the dataset has higher registration accuracy compared with other image registration methods. Meanwhile, the proposed method is compared with state-of-the-art image fusion methods on this dataset. It can be seen from the objective evaluation indicators that the proposed method is 29.03% higher than the comparison method in terms of the accuracy of highlighting thermal anomalies, and the average running time is 2.696 s, which can meet the requirements of auxiliary thermal anomaly detection.

A Model for Expressing Industrial Information Based on Object-Oriented Industrial Heat Sources Detected Using Multi-Source Thermal Anomaly Data in China

Industrial heat sources have made a great contribution to Chinese economic development. However, it has also been found that emissions from industrial heat sources are the main contribution to regional air pollution. Therefore, the detection of industrial heat sources and the expression of related information is becoming important. In this paper, the detection of industrial heat sources was used to express industrial information, thus that the accuracy of the detection of industrial thermal anomalies could be improved and the problems of noise and missing parameters addressed. A model for expressing industrial information based on object-oriented industrial heat sources and using multi-source thermal anomaly data in China was, therefore, proposed. It was a new real-time, objective, and real way to describe the production operation status of industrial heat sources on a large-scale area. First, 4340 working industrial heat sources in mainland China were detected by applying an adaptive k-means algorithm to ACF (NPP VIIRS 375-m active fire/hotspot data) data from the period 19 January 2012 to 31 December 2020. Secondly, several features of working industrial heat sources were extracted from NPP VIIRS 375-m active fire/hotspot data (ACF), VIIRS Nightfire data (VNF), and the Fires product based on Landsat-8 AIRCAS (L8F) data. Areas containing working industrial heat sources were then identified based on these different types of fire data. Light, land-surface temperature, and CO2 and N2O emissions data related to the working industrial heat sources were also extracted. The results show that feature parameters extracted from the multi-source thermal anomaly data mostly have a good positive correlation with the other parameters.

Machine learning-based thermal anomalies detection from MODIS LST associated with the Mw 7.7 Awaran, Pakistan earthquake

Machine learning-based thermal anomalies detection from MODIS LST associated with the Mw 7.7 Awaran, Pakistan earthquake

Updating the Earth remote sensing software for the detection of thermal anomalies

O b j e c t i v e s. The task of improving the software package for detecting thermal anomalies based on meteorologicalsatellite data developed by the unitary enterprise "Geoinformation Systems" was solved.M e t h o d s. In the period from 2015 to the present, the work on practical testing and improvement of the software for natural fires detection has been carried out. For this purpose, satellite images of the territory of Belarus obtained from NOAA series spacecraft were used. Special attention was paid to the problem of improving the accuracy of determining the coordinates of fires and reducing the time required for initial data processing.Re s u l t s. A retrospective analysis of the main stages of improving the software for natural fires detection and obtained during practical tests generalized results are provided. The description of the web service developed on the basis of the software for detecting natural fires is presented.Co n c l u s i o n. The information can be useful for the specialists and researchers who are engaged in the detection of thermal anomalies (fires) using remote sensing data from meteorological satellites.

Detection of thermal anomalies on building façades using infrared thermography and supervised learning

Detection of thermal anomalies on building façades using infrared thermography and supervised learning

Detection of thermal anomalies on building façades using infrared thermography and supervised learning

We propose a cost-effective, non-intrusive approach to assess the quality of the thermal isolation of the building stock in a given municipality or small city. Our prototype measurement system, mounted on a vehicle, acquires geotagged optical and infrared images from the street-side of the buildings. A basic approach to detect gross thermal anomalies, such as thermal bridges, via a binary classifier trained on 2000 carefully labeled infrared images is demonstrated. Thermal anomalies are identified with a precision score of around 89.2 % and 75.6 % recall on a test dataset of 1184 images infrared recorded in Hergiswil (NW). The enabled automated assessment helps to identify retrofitting targets and supports increasing the renovation rate.

THERMAL ANOMALY DETECTION BASED ON SALIENCY ANALYSIS FROM MULTIMODAL IMAGING SOURCES

Abstract. Thermal anomaly detection has an important role in remote sensing. One of the most widely used instruments for this task is a Thermal InfraRed (TIR) camera. In this work, thermal anomaly detection is formulated as a salient region detection, which is motivated by the assumption that a hot region often attracts attention of the human eye in thermal infrared images. Using TIR and optical images together, our working hypothesis is defined in the following manner: a hot region that appears as a salient region only in the TIR image and not in the optical image is a thermal anomaly. This work presents a two-step classification method for thermal anomaly detection based on an information fusion of saliency maps derived from both, TIR and optical images. Information fusion, based on the Dempster-Shafer evidence theory, is used in the first phase to find the location of regions suspected to be thermal anomalies. This classification problem is formulated as a multi-class problem and is carried out in an unsupervised manner on a pixel level. In the following phase, classification is formulated as a binary region-based problem in order to differentiate between normal temperature variations and thermal anomalies, while Random Forest (RF) is chosen as the classifier. In the seconds phase, the classification results from the previous phase are used as features along with temperature information and height details, which are obtained from a Digital Surface Model (DSM). We tested the approach using a dataset, which was collected from a UAV with TIR and optical cameras for monitoring District Heating Systems (DHS). Despite some limitations outlined in the paper, the presented innovative method to identify thermal anomalies has achieved up to 98.7 percent overall accuracy.

Thermal Anomaly Detection of Industrial Zones with MNF and ICA

Thermal anomalies can be detected with the help of the imagery provided by the satellite systems such as Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER). ASTER provides five thermal bands for the effective analysis of thermal anomalies. In order to achieve this goal, considering the physical phenomena, many satellite signal processing methods and algorithms can be used. In this study, depending on the studied area, heat characteristics and extent are presented by using four days of data from daytime and nighttime scenes. In order to define the thermal anomalies for the studied area, Land Surface Temperature (LST) was estimated by inverse Planck function approach for all TIR bands. Minimum Noise Fraction (MNF) and Independent Component Analysis (ICA) methods were applied on all thermal infrared (TIR) bands. The results of MNF and ICA components show location of the thermal anomalies for industrial complexes especially in nighttime scenes.

Model-based thermal anomaly detection for lithium-ion batteries using multiple-model residual generation

The continuously increasing energy and power density of lithium-ion batteries will aggravate the safety and reliability concerns of advanced battery management systems (BMSs). To ensure the safety and reliability of lithium-ion batteries, the BMS must implement anomaly detection algorithms that are capable of capturing abnormal behaviors. Thermal anomalies are one of the most critical anomalies that can be potentially catastrophic. Motivated by this, a model-based strategy of anomaly detection of thermal parameters for lithium-ion-batteries is presented in this paper. The algorithm is based on a multiple-model adaptive estimation framework. Firstly, an equivalent-circuit-model-based electrothermal model is proposed to describe battery dynamic behaviors. Then, a combination of the recursive-least-square method and Kalman-filter is employed to generate residual signals for thermal anomaly detection. Furthermore, the probability of the signature anomaly is evaluated through the multiple-model adaptive estimation technique. Distinguished from existing threshold-based methods, the proposed method can determine particular anomalies according to the value of the generated conditional probability, without a manually determined threshold. Simulations are developed to simulate different faults and generate data for algorithm validation. The results show signature thermal anomaly can be detected accurately.

A computer system for searching and evaluating thermal anomalies in volcano images

The article discusses the analysis of images obtained from video cameras shooting in a wide range, including visible and near-infrared wavelengths. The problem of identification of thermal anomalies in the images of volcanoes obtained with such cameras is solved. The introduction contains a brief review of methods and approaches to the solution of the problem. Section 1 is devoted to the description of the developed algorithm of thermal anomalies detection. At the beginning of the section, the procedure of data sampling preparation is described and the details of thermal anomalies display on the images are shown. Then a description of all the stages of the algorithm is given, including finding the centers of potential anomalies, calculating the area and determining the features of the anomalies with further classification of the obtained data into classes of "thermal" anomaly and "non-thermal" anomaly. In Section 2, the description of the computer system for automated image analysis is given. At the beginning, the implementation of the developed algorithm in the form of a console application is shown. Then its integration into the computer system for the automated analysis of images of the system of continuous video observation of volcanoes of Kamchatka is considered. Also is given a description of all capabilities of the obtained system, a scheme of its work and examples of developed interfaces. In the conclusion of the article summarized the results of the work and the prospects for further research.