Stationary Wavelet Research Articles

Internal Combustion Engine Fault Detection Based on Random Convolutional Neural Networks

The internal combustion engine plays a very important role in many fields, and when the internal combustion engine fails, if it is not found in time, it may cause continuous damage to the internal combustion engine, and further affect the life of the entire mechanical system. To solve the above problems, this paper proposes fault detection of internal combustion engines based on a random convolutional neural network. The cylinder burst noise signal of the internal combustion engine is decomposed by a stationary wavelet, and the inclusion matrix is composed of partial decomposition coefficients. Using singular value theory, the singular value containing matrix is extracted as the characteristic of cylinder burst noise signal. The singular value is used as the input of a random convolutional neural network to train and identify faults. The AdaBound optimizer was then applied to adapt the learning rate to changes, thereby accelerating the weight update of the model. At the same time, the neurons in the structure are randomly deactivated by dropout technology to prevent complex collaborative responses to the training data, and the diagnosis results of each network model are integrated by Dempster synthesis rules. The experimental results show that the minimum mean square error of 0.00165 is achieved when there are 15 neurons in the hidden layer, and it gradually increases as the number of neurons increases. Therefore, it is determined that there should be 15 neurons in the hidden layer, and the trainLM algorithm is used. The decision factors for training, validation, cross-validation, and overall data are 0.98947, 0.98597, 0.97738, and 0.9801, respectively, all reaching the control accuracy of 0.95, which indicates that the random convolutional neural network proposed has a high accuracy for internal combustion engine fault detection. For the main bearing Y-directional force, as the gap increases, its fluctuation trend strengthens; the main bearing force changes within a small range at first, then increases sharply when the gap reaches 0.2[Formula: see text]mm, and some main bearings even experience significant impact loads.

Turbulence Revealed by Wavelet Transform: Power Spectrum and Intermittency for the Velocity Field of the Cosmic Baryonic Fluid

We use continuous wavelet transform techniques to construct the global and environment-dependent wavelet statistics, such as energy spectrum and kurtosis, to study the fluctuation and intermittency of the turbulent motion in the cosmic fluid velocity field with the IllustrisTNG simulation data. We find that the peak scale of the energy spectrum defines a characteristic scale, which can be regarded as the integral scale of turbulence, and the Nyquist wavenumber can be regarded as the dissipation scale. With these two characteristic scales, the energy spectrum can be divided into the energy-containing range, the inertial range, and the dissipation range of turbulence. The wavelet kurtosis is an increasing function of the wavenumber k, which first grows rapidly then slowly with k, indicating that the cosmic fluid becomes increasingly intermittent with k. In the energy-containing range, the energy spectrum increases significantly from z = 2 to 1, but remains almost unchanged from z = 1 to 0. We find that both the environment-dependent spectrum and kurtosis are similar to the global ones, and the magnitude of the spectrum is smallest in the lowest-density and largest in the highest-density environment, suggesting that the cosmic fluid is more turbulent in a high-density than in a low-density environment. In the inertial range, the energy spectrum’s exponent is steeper than both the Kolmogorov and Burgers exponents, indicating more efficient energy transfer compared to Kolmogorov or Burgers turbulence.

Predicting Power Consumption Using Deep Learning with Stationary Wavelet

Predicting Power Consumption Using Deep Learning with Stationary Wavelet

Temporal assessment of Gross alpha emissions from the petroleum industry in Tenerife, Canary Islands (2001–2022)

A ca. 76% decrease in gross alpha activity levels, measured in surface aerosols collected in the city of Santa Cruz de Tenerife (Spain), has been explained in the present study in connection with the reduction of activities, and eventual closure, of an oil refinery in the city. Gross Alpha in surface aerosols, collected at weekly intervals over a period of 22 years (2001–2022), was used for the analysis. The dynamic behaviour of the gross alpha time series was studied using statistical wavelet, multifractal analysis, empirical decomposition method, multivariate analysis, principal component, and cluster analyses approaches. This was performed to separate the impact of other sources of alpha emitting radionuclides influencing the gross alpha levels at this site. These in-depth analyses revealed a noteworthy shift in the dynamic behaviour of the gross alpha levels following the refinery's closure in 2013. This analysis also attributed fluctuations and trends in the gross alpha levels to factors such as the 2008 global economic crisis and the refinery's gradual reduction of activity leading up to its closure. The mixed-model approach, incorporating multivariate regression and autoregressive integrated moving average methods, explained approximately 84% of the variance of the gross alpha levels. Finally, this work underscored the marked reduction in alpha activity levels following the refinery's closure, alongside the decline of other pollutants (CO, SO2, NO, NO2, Benzene, Toluene and Xylene) linked to the primary industrial activity in the municipality of Santa Cruz de Tenerife.

Wavelet-based statistics for enhanced 21cm EoR parameter constraints

We propose a new approach to improve the precision of astrophysical parameter constraints for the 21cm signal from the epoch of reionisation (EoR). Our method introduces new sets of summary statistics, hereafter ‘evolution compressed’ statistics, which quantify the spectral evolution of the 2D spatial statistics computed at fixed redshift. We defined such compressed statistics for power spectrum (PS), wavelet scattering transforms (WST), and wavelet moments (WM), which also characterise non-Gaussian features. To compare these different statistics, along with the 3D power spectrum, we estimated their Fisher information on three cosmological parameters from an ensemble of simulations of 21cm EoR data, both in noiseless and noisy scenarios using Square Kilometre Array (SKA) noise levels equivalent to 100 and 1000 h of observations. We also compare wavelet statistics, in particular WST, built from standard directional Morlet wavelets, as well as from a set of isotropic wavelets derived from the binning window function of the 2D power spectrum. For the noiseless case, the compressed wavelet statistics give constraints that are up to five times more precise than those obtained from the 3D isotropic power spectrum. At the same time, for 100 h SKA noise, from which it is difficult to extract non-Gaussian features, compressed wavelet statistics still give over 30% tighter constraints. We find that the wavelet statistics with wavelets derived from the power-spectrum binning window function provide the tightest constraints of all the statistics, with the WSTs seemingly performing better than the WMs, in particular when working with noisy data. The findings of this study demonstrate that evolution-compressed statistics extract more information than usual 3D isotropic power-spectra approaches and that our wavelet-based statistics can consistently outmatch power-spectrum-based statistics. When constructing such wavelet-based statistics, we also emphasise the need to choose a set of wavelets with an appropriate spectral resolution concerning the astrophysical process studied.

Video source camera identification using fusion of texture features and noise fingerprint

In Video forensics, the objective of Source Camera Identification (SCI) is to identify and verify the origin of a video that is under investigation. This aids the investigator to trace the video to its owner or narrow down the search space for identifying the offender. Nowadays, it is easy to record and share videos via internet or social media with smartphones. The availability of sophisticated video editing tools and software allow offenders to modify video's context. Thus, identifying the right source camera that was used to capture the video becomes complicated and strenuous. Existing methods based on video metadata information are no longer reliable as it could be modified or stripped off. Better forensic procedures are therefore required to prove the authenticity and integrity of the video that will be used as evidence in court of law. Certain inherent camera sensor properties such as, subtle traces of Photo Response Non-Uniformity (PRNU) are present in all captured videos due to unnoticeable defect during the manufacture of camera's sensor. These properties are used in SCI to classify devices or models as they are unique. In this work, we focus on SCI from videos or Video Source Camera Identification (VSCI) to verify the authenticity of videos. PRNU can be affected by highly textured content or post-processing when computed from a set of flat field images. To mitigate these effects, Higher Order Wavelet Statistics (HOWS) information from PRNU of a video I-frame is combined with information from two other texture features i.e., Local Binary Pattern (LBP) and Gray Level Co-occurrence Matrix (GLCM). The extracted feature vector is fused via concatenation and fed to Support Vector Machine (SVM) classifier to perform training and testing for VSCI. Experimental evaluation of our proposed method on videos from different publicly available datasets show the effectiveness of our method in terms of accuracy, resource efficiency, and complexity.

STATISTICAL ANALYSIS BY WAVELET LEADERS REVEALS DIFFERENCES IN MULTI-FRACTAL CHARACTERISTICS OF STOCK PRICE AND RETURN SERIES IN TURKISH HIGH FREQUENCY DATA

The price and return time series are two distinct features of any financial asset. Hence, examining the evolution of multiscale characteristics of price and returns sequential data in time domain would be helpful in gaining a better understanding of the dynamical evolution mechanism of the financial asset as a complex system. In fact, this is important to understand their respective dynamics and to design their appropriate predictive models. The main purpose of the current work is to investigate the multiscale fractals of price and return high frequency data in Turkish stock market. In this regard, the wavelet leaders computational method is applied to each high frequency data to reveal its multi-fractal behavior. In particular, the method is applied to a large set of Turkish stocks and statistical results are performed to check for (i) presence of multi-fractals in price and return series and (ii) differences between prices and returns in terms of multi-fractals. Our statistical results show strong evidence that high frequency price and return data exhibit multi-fractal dynamics. In addition, they show evidence of distinct fractal characteristics on different scales between price and return series. Furthermore, our statistical results show evidence of differences in local fluctuation characteristics of price and return time series. Therefore, differences in local characteristics are useful to build specific predictive models for each type of data for better modeling and prediction to generate profits. Besides, we found evidence that both long-range correlations and fat-tail distributions contribute to the multifractality in Turkish stocks. This finding can be attributed to the major role played by international investors in increasing the volatility of Turkish stocks.

Diagnosing Cracks in the Injector Nozzles of Marine Internal Combustion Engines during Operation Using Vibration Symptoms

In the operation of internal combustion engines, despite technical state monitoring, some cracks that develop in metal components go undetected, leading to secondary, critical, or degradation damage. The diagnostic systems used in floating objects mainly use quasi-static thermodynamic signals, which alert operators too late about emerging damage. Although various methods have been developed to detect cracks in internal combustion engine components, the effectiveness and implementation of the proposed methods are not satisfactory. Therefore, this article presents the use of selected vibration and in-cylinder pressure signals to diagnose the development of damage in some components of marine diesel engines. The investigations were conducted under the natural conditions of the operation of sea-going vessels during port-handling operations. During these investigations, it was possible to observe clear changes in the values of diagnostic symptoms, which corresponded to the development of damage. The developing damage detected in the study involved cracks in injector nozzles manufactured from alloy steel. Despite advances in design, materials, and manufacturing technology, injector nozzle cracks still occur. The diagnostic symptoms used to detect damage development were the amplitude and spectral and wavelet measurements of vibration acceleration signals. This work aimed to search for crack-oriented methods of signal analysis, for example, computer visualization and the recording of diagnostic parameters in various domains. Decimation, windowed, time, amplitude, and time-frequency domain analyses; wavelet statistics; color analysis; and machine learning were used for classification using artificial neural networks. Experimental investigations showed the possibility of diagnosing the development processes of damage to marine diesel engines. The advanced signal processing methods used made it possible to obtain many signal measurements, from which the most useful diagnostic symptoms were selected. The new symptoms found with decimation, time-domain windowed analysis, and Haar wavelet statistics were more useful than the existing ones.

A high-resolution method of spectral fitting with horizon constraints and its application

Abstract When seismic waves propagate underground, subsurface media can absorb high-frequency components of the seismic waves, resulting in attenuation of the high-frequency portion of the seismic data. Therefore, the seismic resolution is low and it is difficult to provide the information needed for thin layer prediction. In this paper, a spectral fitting method with horizon constraints is proposed. This method fully takes into account the spectral characteristics and tectonic changes of the seismic data. It can effectively extend the seismic frequency bandwidth by spectral fitting and improve the seismic resolution. At the same time, the structural information is integrated into the high-resolution processing as the layer information is used to constrain the target equation, and more accurate spectral features of multi-channel data can be obtained. Then the spectral weighting coefficient can be calculated more accurately, resulting in more realistic and accurate seismic data. At the same time, the seismic phase is not destroyed by the processing, so the structural features become clearer, especially for small fractures and thin layers. This method is used for comparison with traditional robust deconvolution and statistical wavelet deconvolution. The spectral components are more faithful and the resolution is higher when processed with this method.

Single Image based Super Resolution Ultrasound Imaging Using Residual Learning of Wavelet Features.

The generation of super resolution ultrasound images from the low-resolution (LR) brightness mode (B-mode) images acquired by the portable point of care ultrasound systems has been of sufficient interest in the recent past. With the advancements in deep learning, there have been numerous attempts in this direction. However, all the approaches have been concentrated on employing the direct image as the input to the neural network. In this work, a stationary wavelet (SWT) decomposition is employed to extract the features from the input LR image which is passed through a modified residual network and the learned features are combined using the inverse SWT to reconstruct the high resolution (HR) image at a 4× scale factor. The proposed approach when compared to the state-of-the art approaches, results in an improved high resolution reconstruction.Clinical relevance- The proposed approach will enable the generation of high-resolution images from portable ultrasound systems, allowing for easier interpretation and faster diagnostics in primary care settings.

Log-constrained inversion based on a conjugate gradient-particle swarm optimization algorithm

Well-logging-constrained impedance inversion is an effective process for predicting the thickness and bifurcation of coal seams. Wavelet changes in a complex region achieve the best match between the inverse and source wavelets, affecting the accuracy of the inversion solution and the ability to obtain accurate inverted acoustic impedance (AI) data. We have conducted the joint inversion of wavelet and AI data using iterative methods, which combined the conjugate-gradient (CG) method and particle-swarm-optimization (PSO) algorithm. The Marmousi AI model was used to prove the reliability of the method. The CG-PSO algorithm achieved excellent results compared with the statistical wavelet pickup method. The wavelet obtained by the CG-PSO algorithm is preferred for inversion operations. We applied a new method to invert field data and predict the thickness and bifurcation of coal seams in the karst region. The results find that the wavelet spectrum obtained by the CG-PSO matches the spectrum map of the coal seam in the Yuwang colliery. We determined the distribution of the thickness and bifurcation of the 101 panel, Yuwang Colliery, Yunnan Province. The average error of the predicted coal thickness is 0.17 m (14.4%), which verifies the feasibility and effectiveness of the method. The method provides insights into the AI inversion of constrained waves in complex regions.

SWM-DE: Statistical wavelet model for joint denoising and enhancement for multimodal medical images

Medical images are usually degraded by numerous noises during acquisition or transmission, which often causes low contrast leading to deterioration of image quality. As such, medical image denoising and enhancement has become a paramount routine task. To overcome this problem, we propose a cutting-edge joint statistical and morphological model for the denoising and enhancement operation. Firstly, we propose a statistical model in formulating the marginal distribution of the wavelet coefficients. This model is integrated into a Bayesian inference framework to develop a maximum a posterior (MAP) estimator of the noise-free coefficient. Based on the statistical model, we eliminate the need for noise level estimation, and allows the model to automatically adapts to the observed image data. Secondly, we propose an adjustable morphological reconstruction model to eliminate known and unknown noises associated with medical images, while preserving the image details. After these operations, the image is decomposed into several wavelet subbands to extract the illumination and detail components. The image is then reconstructed based on the inverse wavelet to generate the enhanced noise-free image.Experimental results show that the proposed framework obtained high EME values of 41.04, 48.81, 47.81, and 45.75 for OCTA, FFA, CT, and X-ray imaging modalities, and performs better than the state-of-the-art methods. The proposed algorithm can effectively and efficiently enhance medical images, which will assist the clinicians in disease diagnosis, monitoring, and treatment.

SimBIG: mock challenge for a forward modeling approach to galaxy clustering

Simulation-Based Inference of Galaxies (SimBIG) is a forward modeling framework for analyzing galaxy clustering using simulation-based inference. In this work, we present the SimBIG forward model, which is designed to match the observed SDSS-III BOSS CMASS galaxy sample. The forward model is based on high-resolution Quijote N-body simulations and a flexible halo occupation model. It includes full survey realism and models observational systematics such as angular masking and fiber collisions. We present the “mock challenge” for validating the accuracy of posteriors inferred from SimBIG using a suite of 1,500 test simulations constructed using forward models with a different N-body simulation, halo finder, and halo occupation prescription. As a demonstration of SimBIG, we analyze the power spectrum multipoles out to k max = 0.5 h/Mpc and infer the posterior of ΛCDM cosmological and halo occupation parameters. Based on the mock challenge, we find that our constraints on Ω m and σ 8 are unbiased, but conservative. Hence, the mock challenge demonstrates that SimBIG provides a robust framework for inferring cosmological parameters from galaxy clustering on non-linear scales and a complete framework for handling observational systematics. In subsequent work, we will use SimBIG to analyze summary statistics beyond the power spectrum including the bispectrum, marked power spectrum, skew spectrum, wavelet statistics, and field-level statistics.

Automating seismic‐well tie via self‐supervised learning

AbstractAs an essential process in subsurface interpretation, seismic‐well tie aims at calibrating the well measurements in depth domain with the seismic records in time domain for reliable reservoir mapping and modelling. Such a process usually requires massive manual efforts, primarily extracting source wavelets to synthesize seismograms from well measurements and stretching/squeezing synthetic seismograms to match actual seismic signals, and thus becomes time‐consuming and labour intensive with the amount of wells expanding in a field. This paper formulates the seismic‐well tie problem from the perspective of computer vision and, considering the lack of manually prepared training labels in most of real projects, proposes automating it by a self‐supervised workflow of two key components. The first component is to extract time‐variant wavelets in the target interval using a dual‐task autoencoder, which is optimized by maximizing the spectrum similarity between the extracted wavelet and the actual seismic. The second component is to estimate the corresponding time‐shift between a pair of the synthetic seismogram and the actual seismic using a flow net, which is trained by a pseudo dataset derived from the actual well measurements. More specifically, the data generation consists with defining one‐dimensional time‐shift curves to warp the original less accurate time‐depth relationships and then synthesizing the corresponding seismograms based on the convolutional model at all available wells. When turning to the stage of inference at a target well, the proposed workflow automatically extracts a representative wavelet, generates the corresponding synthetic seismogram from the original time‐depth relationship and estimates the time‐shift curve necessary for revising the original time‐depth relationship that would lead to an optimal tying with the actual seismic at the well. The proposed workflow is tested on two field datasets, and both results demonstrate improved tying over traditional methods such as statistical wavelet extraction and dynamic time warping.

Scale‐separation diagnostics and the Symmetric Bounded Efficiency for the inter‐comparison of precipitation reanalyses

AbstractThe ERA5 global reanalysis has been compared against a high‐resolution regional reanalysis (COSMO‐REA6) by means of scale‐separation diagnostics based on 2d Haar discrete wavelet transforms. The presented method builds upon existing methods and enables the assessment of bias, error and skill for individual spatial scales, separately. A new skill score (evaluated against random chance) and the Symmetric Bounded Efficiency are introduced. These are compared to the Nash‐Sutcliffe and the Kling‐Gupta Efficiencies, evaluated on different scales, and the benefits of symmetric statistics are illustrated. As expected, the wavelet statistics show that the coarser resolution ERA5 products underestimate small‐to‐medium scale precipitation compared to COSMO‐REA6. The newly introduced skill score shows that the ERA5 control member (EA‐HRES), despite its higher variability, exhibits better skill in representing small‐to‐medium scales with respect to the smoother ensemble members. The Symmetric Bounded Efficiency is suitable for the inter‐comparison of reanalyses, since it is invariant with respect to the order of comparison.

Screening Lung Diseases Using Cascaded Feature Generation and Selection Strategies.

The global pandemic COVID-19 is still a cause of a health emergency in several parts of the world. Apart from standard testing techniques to identify positive cases, auxiliary tools based on artificial intelligence can help with the identification and containment of the disease. The need for the development of alternative smart diagnostic tools to combat the COVID-19 pandemic has become more urgent. In this study, a smart auxiliary framework based on machine learning (ML) is proposed; it can help medical practitioners in the identification of COVID-19-affected patients, among others with pneumonia and healthy individuals, and can help in monitoring the status of COVID-19 cases using X-ray images. We investigated the application of transfer-learning (TL) networks and various feature-selection techniques for improving the classification accuracy of ML classifiers. Three different TL networks were tested to generate relevant features from images; these TL networks include AlexNet, ResNet101, and SqueezeNet. The generated relevant features were further refined by applying feature-selection methods that include iterative neighborhood component analysis (iNCA), iterative chi-square (iChi2), and iterative maximum relevance–minimum redundancy (iMRMR). Finally, classification was performed using convolutional neural network (CNN), linear discriminant analysis (LDA), and support vector machine (SVM) classifiers. Moreover, the study exploited stationary wavelet (SW) transform to handle the overfitting problem by decomposing each image in the training set up to three levels. Furthermore, it enhanced the dataset, using various operations as data-augmentation techniques, including random rotation, translation, and shear operations. The analysis revealed that the combination of AlexNet, ResNet101, SqueezeNet, iChi2, and SVM was very effective in the classification of X-ray images, producing a classification accuracy of 99.2%. Similarly, AlexNet, ResNet101, and SqueezeNet, along with iChi2 and the proposed CNN network, yielded 99.0% accuracy. The results showed that the cascaded feature generator and selection strategies significantly affected the performance accuracy of the classifier.

Experimental investigations on efficiency and instability of combustion process in a diesel engine fueled with ternary blends of hydrogen peroxide additive/biodiesel/diesel

ABSTRACT The current study intends to assess the impact of incorporating an oxygenated fuel additive (hydrogen peroxide (H2O2)) into a biodiesel/diesel fuel mix on diesel engine combustion efficiency and instability. The combustion, performance, and emission properties were evaluated using advanced test equipment to offer a full study. The combustion instabilities in terms of cyclic fluctuations were investigated to discover any unusual behavior when H2O2 was added to biodiesel/diesel fuel mixtures. Statistical and continuous wavelet transformations were used to investigate the cyclic fluctuations in peak cylinder pressure and indicated mean effective pressure (IMEP). Diesel, B20 (20%/80% of biodiesel/diesel fuel), B20H5 (20%/5%/75% of biodiesel/H2O2/diesel fuel), and B20H10 (20%/10%/70% of biodiesel/H2O2/diesel fuel) were tested in the study. Experiment results indicated that B20H10 produced the best results. In comparison to B20, B20H10 improved BTE by 6.56% and HRR by 8.5%, while decreasing BSFC by 7.98%, CO and HC emissions by 26.5%, and 6.67%, respectively. Furthermore, the peak pressure and IMEP cyclic changes were also well within permissible ranges. However, it was shown that adding H2O2 to B20 increased NOx emissions. The cyclic variations in the cylinder peak pressure and indicated mean effective pressure were analyzed using statistical methods and continuous wavelet transformation. No abnormality was detected in cyclic variation, thus ruling out any unsafe behavior in introducing test fuel additive to biodiesel blends. In principle, introducing H2O2 into a biodiesel/diesel fuel mix helped improve the engine performance and emissions while having no negative effects on engine life.

Machine learning-based vertical resolution enhancement considering the seismic attenuation

The vertical resolution of seismic data is important for the interpretation of geologic features at a fine scale. To improve the resolution of seismic data, spectral enhancement has progressed mainly due to one-dimensional (1D) convolutional models and deconvolution filters; however, recent studies have applied machine learning (ML) algorithms. To generate successful outcome using ML techniques, it is important to reflect the features of target data in the training stage of the ML model. One of main features of seismic field data is a non-stationary wavelet over time due to the change of frequency content as the wavelet passes through specific regions, such as a gas reservoir. However, since conventional methods are generally based on a 1D convolutional model, which assumes that the propagating wavelet is stationary, they show low performance for the parts where the frequency content of the wavelet is significantly changed. In this study, we developed a spectral enhancement algorithm using ML techniques, which can be applied to a seismic trace with different frequency contents over time. We showed that sparse spike inversion results of all target data, rather than just well log data, can be useful to obtain information on the reflectivity series for generating a training data set. To verify the performance of the new algorithm, we compared the spectrally enhanced results from the ML model trained by stationary wavelets and time-variant wavelets and confirmed that the latter wavelets provided better results. In addition, we proposed a quality control (QC) method for verifying the spectrally enhanced results. • We developed a spectral enhancement ML algorithm consider the seismic attenuation. • The ML model trained by time-variant wavelets yielded better results for the area with seismic attenuation. • The results obtained by the ML were verified the validity by our proposed QC method only using seismic data.

Multivariate Statistical Modeling for Multitemporal SAR Change Detection Using Wavelet Transforms and Integrating Subband Dependencies

In this paper, we propose a new method for automatic change detection in multi-temporal fully polarimetric synthetic aperture radar (PolSAR) images based on multivariate statistical wavelet subband modeling. The proposed method allows us to take into account the correlation structure between subbands by modeling the wavelet coefficients through multi-variate probability distributions. Three types of correlation are investigated: inter-scale, inter-orientation, and inter-polarization dependences. The multivariate generalized Gaussian distribution (MGGD) is used to model the interdependencies between wavelet coefficients at different orientations, scales, and polarizations. Kullback-Leibler similarity measures are computed and used to generate the change map. Simulated and real multilook PolSAR data are employed to assess the performance of the method and are compared to the multivariate Gaussian distribution (MGD) based method. We show that the information embedded in the correlation between subbands improves the accuracy of the change map, leading to better performance. Moreover, the MGGD represents better the correlations between wavelet coefficients and outperforms the MGD.

Thermal Imagery Feature Extraction Techniques and the Effects on Machine Learning Models for Smart HVAC Efficiency in Building Energy

The control of thermostats of a heating, ventilation, and air-conditioning (HVAC) system installed in commercial and residential buildings remains a pertinent problem in building energy efficiency and thermal comfort research. The ability to determine the number of people at a particular time in an area is imperative for energy efficiency in order to condition only occupied regions and thermally deficient regions. In this study of the best features comparison for detecting the number of people in an area, feature extraction techniques including wavelet scattering, wavelet decomposition, grey-level co-occurrence matrix (GLCM) and feature maps convolution neural network (CNN) layers were explored using thermal camera imagery. Specifically, the pretrained CNN networks explored are the deep residual (Resnet-50) and visual geometry group (VGG-16) networks. The discriminating potential of Haar, Daubechies and Symlets wavelet statistics on different distributions of data were investigated. The performance of VGG-16 and ResNet-50 in an end-to-end manner utilizing transfer learning approach was investigated. Experimental results showed the classification and regression trees (CART) model trained on only GLCM and Haar wavelet statistic features, individually achieved accuracies of approximately 80% and 84%, respectively, in the detection problem. Moreover, k-nearest neighbors (KNN) trained on the combined features of GLCM and Haar wavelet statistics achieved an accuracy of approximately 86%. In addition, the performance accuracy of the multi classification support vector machine (SVM) trained on deep features obtained from layers of pretrained ResNet-50 and VGG-16 was between 96% and 97%. Furthermore, ResNet-50 transfer learning outperformed the VGG-16 transfer learning model for occupancy detection using thermal imagery. Overall, the SVM model trained on features extracted from wavelet scattering emerged as the best performing classifier with an accuracy of 100%. A principal component analysis (PCA) on the wavelet scattering features proved that the first twenty (20) principal components achieved a similar accuracy level instead of training on the whole feature set to reduce the execution time. The occupancy detection models can be integrated into HVAC control systems for energy efficiency and security systems, and aid in the distribution of resources to people in an area.