Surface Anomaly Research Articles

Multi-Scale Rail Surface Anomaly Detection Based on Weighted Multivariate Gaussian Distribution

Multi-Scale Rail Surface Anomaly Detection Based on Weighted Multivariate Gaussian Distribution

Vision-Language Interaction via Contrastive Learning for Surface Anomaly Detection in Consumer Electronics Manufacturing

Vision-Language Interaction via Contrastive Learning for Surface Anomaly Detection in Consumer Electronics Manufacturing

Toward a Novel Method for General On-Orbit Earth Surface Anomaly Detection Leveraging Large Vision Models and Lightweight Priors

Toward a Novel Method for General On-Orbit Earth Surface Anomaly Detection Leveraging Large Vision Models and Lightweight Priors

A deep learning approach for anomaly identification in PZT sensors using point contact method

The implementation of piezoelectric sensors is degraded due to surface defects, delamination, and extreme weathering conditions, to mention a few. Hence, the sensor needs to be diagnosed before the efficacious implementation in the structural health monitoring (SHM) framework. To rescue the problem, a novel experimental method based on Coulomb coupling is utilised to visualise the evolution of elastic waves and interaction with the surface anomaly in the lead zirconate titanate (PZT) substrate. Recently, machine learning (ML) has been expeditiously becoming an essential technology for scientific computing, with several possibilities to advance the field of SHM. This study employs a deep learning-based autoencoder neural network in conjunction with image registration and peak signal-to-noise ratio (PSNR) to diagnose the surface anomaly in the PZT substrate. The autoencoder extracts the significant damage-sensitive features from the complex waveform big data. Further, it provides a nonlinear input–output model that is well suited for the non-linear interaction of the wave with the surface anomaly and boundary of the substrate. The measured time-series waveform data is provided as input into the autoencoder network. The mean absolute error (MAE) between the input and output of the deep learning model is evaluated to detect the anomaly. The MAEs are sensitive to the anomaly that lies in the PZT substrate. Further, the challenge arising from offset and distortion is addressed with ad hoc image registration technique. Finally, the localisation and quantification of the anomaly are performed by computing PSNR values. This work proposes an advanced, efficient damage detection algorithm in the scenario of big data that is ubiquitous in SHM.

Real-time remote sensing detection framework of the earth's surface anomalies based on a priori knowledge base

Due to the suddenness, diversity, randomness and complexity of various earth's surface anomalies caused by natural and human factors, the time-efficiency of traditional satellite remote sensing technology is seriously lagging behind the actual needs of the earth's surface anomalies early warning and emergency handling. Real-time earth's surface anomaly detection has become an important issue for achieving ecological security, human safety and health, and high-quality socioeconomic development. Therefore, this study proposed a real-time detection framework for earth's surface anomalies using a single temporal remote sensing data based on a priori knowledge base. The framework first selected an appropriate remote sensing feature system and constructed a priori knowledge base effectively reflecting the natural evolutionary law of the normal earth's surface. Then, the earth's surface anomalies were identified by comparing the real-time acquired remote sensing data to the priori knowledge base with the designed detection rules. Finally, the intensity of earth's surface anomalies was evaluated comprehensively using pixel- and patch-scale information. The preliminary analysis of several specific cases indicated that the proposed framework was reliable and could effectively identify the earth's surface anomalies with high accuracy. This proposed framework is fully applicable to the real-time detection of most earth's surface anomalies using a single temporal remote sensing data, which is not limited to a single earth's surface anomaly event. Meanwhile, the framework presents great potential for on-orbit processing of the earth's surface anomaly detection and realizing real-time remote sensing services for the early warning of the earth's surface anomalies.

NFCF: Industrial Surface Anomaly Detection with Normalizing Flow Cross-Fitting Network

The effective fully-supervised defect detection methods in industry are done by training many defect labels. These methods face application challenges because of insufficient defect samples. Moreover, the subtlety of defects, the similarity between defect-free and defect regions, and the interference factors carried by defect-free images also pose problems. To solve these difficulties above, we propose a Normalizing Flow Cross-Fitting network (NFCF) to use only defect-free images as priori knowledge for training. Specifically, the incremental broadening module in the network focuses on information expansion for minor defects, and the interactive filtering module completes the weight filtering through same-scale mutual reasoning. The processed information is fitted to a defect-free distribution by the normalizing flow module and the distribution is used as the basis for the determination. The NFCF does experiments on four types of datasets, and it achieves an average of 96.3% and 95.88% on two metrics, AUC-Image and AUC-Pixel. Experimental results show that the network can distinguish defect images using only defect-free training. Moreover, visualization experiments show that it can also complete localization segmentation for minor defects. Based on semi-supervision with achieved accuracy, NFCF alleviates the labeling dependence problem of defects and significantly increases the application value.

Detection of Road Surface Anomaly Using Distributed Fiber Optic Sensing

Road surface condition can significantly impact the interaction between vehicles and pavement structure, which may even cause high fuel consumption and safety issues of drivers and vehicles. Distributed fiber optic sensing (DFOS) technology is a useful tool to perform continuous and real-time monitoring of traffic and road surface condition. However, it is challenging to process the data for the purpose of road anomaly detection. The study proposed two approaches to detect the road anomaly using DFOS. In the first method, local binary pattern (LBP) histograms were used to extract the features of the images with and without road anomaly, and support vector machine (SVM) combined with principal component analysis (PCA) was adopted as the classifier. The convolutional neural network (CNN) was applied on the binary classification data to analyze the images in the second method. The accuracy and benefits of two methodologies were compared. The vehicle speed was estimated by detecting lines using Hough transform. The feasibility of road anomaly detection using DFOS is proved.

A Review of Road Surface Anomaly Detection and Classification Systems Based on Vibration-Based Techniques

Road surfaces suffer from sources of deterioration, such as weather conditions, constant usage, loads, and the age of the infrastructure. These sources of decay generate anomalies that could cause harm to vehicle users and pedestrians and also develop a high cost to repair the irregularities. These drawbacks have motivated the development of systems that automatically detect and classify road anomalies. This study presents a narrative review focused on road surface anomaly detection and classification based on vibration-based techniques. Three methodologies were surveyed: threshold-based methods, feature extraction techniques, and deep learning techniques. Furthermore, datasets, signals, preprocessing steps, and feature extraction techniques are also presented. The results of this review show that road surface anomaly detection and classification performed through vibration-based methods have achieved relatively high performance. However, there are challenges related to the reproduction and heterogeneity of the results that have been reported that are influenced by the limited testing conditions, sample size, and lack of publicly available datasets. Finally, there is potential to standardize the features computed through the time or frequency domains and evaluate and compare the diverse set of settings of time-frequency methods used for feature extraction and signal representation.

Airborne hyperspectroscopy using AVIRIS- NG data for detecting surface signatures of kimberlite and kimberlite clan of rocks in Bundelkhand Craton, India

Abstract We have utilized spectral bands of Advanced Visible Infrared Imaging Spectroradiometer Next Generation sensor (AVIRIS-NG) to identify surface signatures of new kimberlite. The main focus of the work is to identify the surface signatures of probable kimberlite occurrence and to substantiate these surface signatures as the prospective areas of detailed exploration. Since, all the discovered kimberlites were reported within ferruginous quartzite (Fe-quartzite), we prepared an AVIRIS-NG image composite to demarcate the spatial extension of ferruginous quartziteThe AVIRIS-NG image spectrum of a known kimberlite was also used, as a reference, to identify the surface signatures of new kimberlite within Fe-quartzite. We applied matched filtering and adaptive coherence estimator algorithms to detect surface signatures of kimberlite. The detected surface anomalies of kimberlite in each spectral map are consistent in shape, number and spatial extent. Subsequently, we integrated these maps into a unified surface anomaly map of kimberlite using multiplication operator for identifying the surface anomalies of kimberlite. We observed that the image spectra of these surface anomalies are similar to the laboratory and image spectra of a known pipe. Most of the detected new anomalies are circular and associated with geological lineaments delineated in the residual gravity anomaly map derived from the EIGEN 6C4 gravity model. The surface anomalies of kimberlite, identified in this research, substantiate the importance of detected anomalies for further exploration.

Blinking rate with connected glasses in non‐dry eye and dry eye anomalies in healthy volunteers

AbstractPurposeBlinking rate is an essential physiopathological mechanism aiming to maintain a quality tear film layer for ocular surface. The blinking rate may vary with age, environment, medications, ocular surface conditions or anomalies and individuals. Various studies demonstrated how dry eye the most common ocular surface anomaly affects the blinking rate. As we know dry eye is a common pathology in general population, we included healthy volunteers. Our aim is to study blinking rate with connected glasses in healthy volunteers with dry eye and non‐dry eye anomalies.MethodsIn our study, we selected and included 20 healthy volunteers with 40 eyes (of 25) with signed informed consent, from 10 to 40 yo (mean age 25,5 yo) assessing some of the volunteers may present dry eye anomalies. A signed consent was obtained according to legal regulation. A complete ocular examination was performed including automated NIBUT and Meibography with Lacrydiag (Quantel). The volunteers accepted to wear the connected glasses “Prudensee” (Ellcie Healthy) during one hour in practical day life. Then the collected data were analyzed with an algorithm. The connected glasses integrate an infrared oculometer that measures blinking rate. They also integrate a gyroscope and an accelerometer useful for head speed however mt used in the study.ResultsOf the 20 subjects, 9 were diagnosed with mild to moderate dry eye anomalies. The blinking rate was significantly increased of 36% comparing to the 11 subjects with no ocular surface anomalies. Blinking rate measured was increased statistically significative in the 3 groups of age 10–20 yo, 20–30 yo and 30–40 yo and a marked trend in dry eye subjects.ConclusionsOur study demonstrated an increase in blinking rate in subjects with dry eye anomalies. Further studies with connected glasses are needed to evaluate blinking rate in various anomalies of ocular surface with larger number of subjects.

Large eddy simulation of boundary-layer turbulence over the heterogeneous surface in the source region of the Yellow River

Abstract. Lake breezes are proved by downdrafts and the divergence flows of zonal wind in the source region of the Yellow River (SRYR) in the daytime based on ERA-Interim reanalysis data. In order to depict the effect of the circulations induced by surface anomaly heating (patches) on the boundary-layer turbulence, the UK Met Office Large Eddy Model was used to produce a set of 1D strip-like surface heat flux distributions based on observations, which were obtained by a field campaign in the Ngoring Lake basin in the summer of 2012. The simulations show that for the cases without background wind, patch-induced circulations (SCs) promote the growth of convective boundary layer (CBL), enhance the turbulent kinetic energy (TKE), and then modify the spatial distribution of TKE. Based on phase-averaged analysis, which separates the attribution from the SCs and the background turbulence, the SCs contribute no more than 10 % to the vertical turbulent intensity, but their contributions to the heat flux can be up to 80 %. The thermal internal boundary layer (TIBL) reduces the wind speed and forms the stable stratification, which produces the obvious change of turbulent momentum flux and heat flux over the heterogeneous surfaces. The increased downdrafts, which mainly occur over the lake patches, carry more warm, dry air down from the free atmosphere. The background wind inhibits the SCs and the development of the CBL; it also weakens the patch-induced turbulent intensity, heat flux, and convective intensity.

A Tensor Voting-Based Surface Anomaly Classification Approach by Using 3D Point Cloud Data

Abstract Advanced three-dimensional (3D) scanning technology has been widely used in many industries to collect the massive point cloud data of artifacts for part dimension measurement and shape analysis. Though point cloud data has product surface quality information, it is challenging to conduct effective surface anomaly classification due to the complex data representation, high-dimensionality, and inconsistent size of the 3D point cloud data within each sample. To deal with these challenges, this paper proposes a tensor voting-based approach for anomaly classification of artifact surfaces. A case study based on 3D scanned data obtained from a manufacturing plant shows the effectiveness of the proposed method.

Reconstructing the time-variable sea surface from tide gauge records using optimal data-dependent triangulations

Reconstructions of sea level prior to the satellite altimeter era are usually derived from tide gauge records; however most algorithms for this assume that modes of sea level variability are stationary which is not true over several decades. Here we suggest a method that is based on optimized data-dependent triangulations of the network of gauge stations. Data-dependent triangulations are triangulations of point sets that rely not only on 2D point positions but also on additional data (here: sea surface anomalies). In particular, min-error criteria have been suggested to construct triangulations that approximate a given surface. In this article, we show how data-dependent triangulations with min-error criteria can be used to reconstruct 2D maps of the sea surface anomaly over a longer time period, assuming that anomalies are continuously monitored at a sparse set of stations and, in addition, observations of a control surface is provided over a shorter time period. At the heart of our method is the idea to learn a min-error triangulation based on the control data that is available, and to use the learned triangulation subsequently to compute piece-wise linear surface models for epochs in which only observations from monitoring stations are given. Moreover, we combine our approach of min-error triangulation with k-order Delaunay triangulation to stabilize the triangles geometrically. We show that this approach is in particular advantageous for the reconstruction of the sea surface by combining tide gauge measurements (which are sparse in space but cover a long period back in time) with data of modern satellite altimetry (which have a high spatial resolution but cover only the last decades). We show how to learn a min-error triangulation and a min-error k-order Delaunay triangulation using an exact algorithm based on integer linear programming. We confront our reconstructions against the Delaunay triangulation which had been proposed earlier for sea-surface modeling and find superior quality. With real data for the North Sea we show that the min-error triangulation outperforms the Delaunay method substantially for reconstructions back in time up to 18 years, and the k-order Delaunay min-error triangulation even up to 21 years for k=2. With a running time of less than one second our approach would be applicable to areas with far greater extent than the North Sea.

Developing a near real-time road surface anomaly detection approach for road surface monitoring

Road surface hazards affect the driving safety and comfort of road users. Recently, smartphones and mobile devices equipped with motion sensors such as accelerometers and gyroscope sensors have attracted researchers’ attention for the development of low-cost approaches for road surface monitoring. However, processing smartphone sensors to monitor road surface conditions is technically challenging due to dissimilar sensor properties, different smartphone placement, and also different vehicle mechanical properties. This study aimed to develop a hybrid method using threshold based and Machine Learning approaches for near real-time detection and classification of road surface anomalies using smartphone sensor data with higher-level accuracy. The proposed algorithm has self-adapting and self-updating capabilities to adapt itself to any type of smartphone and the dynamic behaviors of various vehicles and road surface conditions. A prototype is developed using MATLAB and ArcGIS to perform sensor data analysis, geocoding, geo-visualizing, and data querying for performance evaluation.

The variability of sea surface temperature in the Patagonian Shelf Argentina, from 35 years of satellite information

ABSTRACT The aim of the present work is to analyse the variations of Sea Surface Temperature (SST) and Anomaly of Sea Surface Temperature (ASST) using more than thirty years of satellite data. We studied if these variations have possible ocean-atmosphere connections with the Southern Annular Mode (SAM), the Southern Oscillation Index (SOI), and the Oceanic Niño Index (ONI) in a crucial region for fisheries off the Patagonian Shelf. For the analysis, we used SST data collected from 1985 to 2019 with Advanced Very High Resolution Radiometer (AVHRR) and Moderate Resolution Imaging Spectroradiometer (MODIS) instruments. The two databases were merged to obtain 35 years of continuous SST data. The study area was divided into three zones to present different oceanographic and bathymetric characteristics. These zones are located in the gulf, in the middle of the continental shelf, and at the shelf break. For each of the zones, the monthly SST climatology and the ASST series are shown. To define their periodicity, monthly ASST data were analysed spectrally by Fast Fourier Transform (FFT) and Wavelet Transform. Cross-Wavelet Transform (XWT) was applied to analyse the co-variance of ASST with respect to climate indices. Our results show that the isotherm distribution seems to be separated by the 64º W meridian. To the east of 64º W, it follows an oblique direction from southwest to northeast; to the west of 64º W, a more complex behaviour of the SST field seems to be associated with the coastal dynamics. The ASST series shows that the values of negative anomalies are higher than those of the positive ones. Nevertheless, positive ASSTs are more extreme and persistent in time than the negative ones. The ASST wavelet diagrams denote a common and non-stationary periodicity in the three zones, with statistically significant power spectra in the intra-seasonal (9-month) and inter-annual (24-month and 4-year) time scales. There are clearly common features in the wavelet power of the time series related to ASST and climate forcing. The SAM exhibited the closest relationship with ASST in the three zones. The SAM and ASST indices denote significant power peaks in the 2-year and 4- to 6-year bands.

Presence versus absence of two-dimensional Fermi surface anomalies

We theoretically consider Fermi surface anomalies manifesting in the temperature-dependent quasiparticle properties of two-dimensional (2D) interacting electron systems, comparing and contrasting with the corresponding three-dimensional (3D) Fermi liquid situation. In particular, employing microscopic many-body perturbative techniques, we obtain analytically the leading-order and the next-to-leading-order interaction corrections to the renormalized effective mass for three distinct physical interaction models: electron-phonon, electron-paramagnon, and electron-electron Coulomb coupling. We find that the 2D renormalized effective mass does not develop any Fermi surface anomaly due to electron-phonon interaction, manifesting $\mathcal{O}({T}^{2})$ temperature correction and thus remaining consistent with the Sommerfeld expansion of the noninteracting Fermi function, in contrast to the corresponding 3D situation where the temperature correction to the renormalized effective mass has the anomalous ${T}^{2}logT$ behavior. In contrast, both electron-paramagnon and electron-electron interactions lead to the anomalous $\mathcal{O}(T)$ corrections to the 2D effective mass renormalization in contrast to ${T}^{2}logT$ behavior in the corresponding 3D interacting systems. We provide detailed analytical results, and comment on the conditions under which a ${T}^{2}logT$ term could possibly arise in the 2D quasiparticle effective mass from electron-phonon interactions. We also compare results for the temperature-dependent specific heat in the interacting 2D and 3D Fermi systems, using the close connection between the effective mass and specific heat.

The Prenatal Diagnosis~ What should we do? What can we do?

Purpose: Prenatal diagnostic accuracy has improved; however, the pediatric surgeon’s role remains unclear. This paper aimed to determine the prenatal diagnoses made and the role of pediatric surgeons. Methods: A 6-year retrospective review of 904 pregnancies managed at our institute was conducted. They were classified as a normal pregnancy (NP, n=194), abnormal pregnancy maternal factor (MF, n=449), or abnormal pregnancy fetal factor (FF, n=261). Results: In the FF group, the identified conditions were twin pregnancies (n=75), intrauterine growth restriction (IUGR) (n=49), breech presentation (n=26), arrested development (n=19), hypoamnion (n=42), fetal distress (n=16), hydramnios (n=10), abnormal heart sounds (n=5), meconium staining (n=5), surface anomaly (n=4), calcification (n=2), fetal hydrops (n=2), fetal death (n=2), bowel dilatation (n=2), abdominal mass (n=1) and diaphragmatic hernia (n=1). Case of twin pregnancies, breech presentation, arrested development, IUGR, hypoamnion, abnormal heart sounds, meconium staining and fetal hydrops did not require surgery. Of the 16 cases of fetal distress, 1 had biliary atresia. Of the 10 cases of hydramnios, 1 had meconium peritonitis. Of the 4 with surface anomalies, 3 had gastroschisis. Of the 2 with calcification, 1 had meconium peritonitis. Of the two fetal deaths, 1 had anal atresia suggesting a chromosomal abnormality. Of the 2 cases of bowel dilatation, 1 had bowel atresia. The abnormal mass was caused by adrenal bleeding. These diagnoses were made at an average of 27.4 gestational weeks; however, 2 cases of gastroschisis, suggesting a body stalk anomaly; diaphragmatic hernia; and brain cysts were diagnosed before 20 weeks and were aborted. After delivery, the mortality rate was 0% in neonates treated by pediatric surgeons and neonatologists. Conclusion: Fetal abnormalities are rare; however, early aggressive management with the cooperation of obstetricians and parents is crucial for pediatric surgeons to minimize the effects of anomalies.

Inversion of magnetic data using deep neural networks

As a novel tool, deep learning is used to solve complex problems in the real world and has been successfully applied to invert for seismic and electromagnetic data. In this study, we propose to use deep neural networks (DNNs) to recover the distribution of the physical properties of buried magnetic orebodies from the surface and airborne magnetic anomaly data. This approach is based on data training instead of prior-knowledge assumptions used in traditional inversion methods. Once our generalized network is established, the computing time to predict new magnetic data using this method can be significantly reduced. By implementing the forward modeling of different types of synthetic physical property models, we obtained enough datasets to train a DNN model so that the network can establish a nonlinear mapping directly from magnetic anomaly data to physical properties. The pre-trained network can be used to estimate the distribution of magnetization intensity from new input magnetic data. Two DNN structures were employed to test the feasibility and generalization of the proposed method by implementing Experiments in serval two-dimensional (2D) synthetic examples. Compared with the conventional method, the predicted distribution of magnetization intensity obtained by using our method is more concentrated and has better resolution to determine the boundary of the magnetic body. In a field example of Galinge iron ore deposits in China, the magnetization distribution of concealed iron orebodies inverted by the proposed approach was in good agreement with that detected by borehole data. DNNs have good nonlinear inversion capability and exhibit some excellent merits of strong learning ability, wide coverage and strong adaptability, which is a considerable application prospect in geophysical exploration.

Segmentation-based deep-learning approach for surface-defect detection

Automated surface-anomaly detection using machine learning has become an interesting and promising area of research, with a very high and direct impact on the application domain of visual inspection. Deep-learning methods have become the most suitable approaches for this task. They allow the inspection system to learn to detect the surface anomaly by simply showing it a number of exemplar images. This paper presents a segmentation-based deep-learning architecture that is designed for the detection and segmentation of surface anomalies and is demonstrated on a specific domain of surface-crack detection. The design of the architecture enables the model to be trained using a small number of samples, which is an important requirement for practical applications. The proposed model is compared with the related deep-learning methods, including the state-of-the-art commercial software, showing that the proposed approach outperforms the related methods on the specific domain of surface-crack detection. The large number of experiments also shed light on the required precision of the annotation, the number of required training samples and on the required computational cost. Experiments are performed on a newly created dataset based on a real-world quality control case and demonstrates that the proposed approach is able to learn on a small number of defected surfaces, using only approximately 25-30 defective training samples, instead of hundreds or thousands, which is usually the case in deep-learning applications. This makes the deep-learning method practical for use in industry where the number of available defective samples is limited. The dataset is also made publicly available to encourage the development and evaluation of new methods for surface-defect detection.

Exploring the stratospheric source of ozone pollution over China during the 2016 Group of Twenty summit

The 2016 Group of Twenty (G20) Hangzhou Summit was concurrently accompanied by extratropical cyclogenesis. To investigate whether the extratropical cyclone exerts any impact on the near surface ozone concentration, the Weather Research Forecast with Chemistry (WRF-Chem) model, with stratospheric passive tracers turned on, was used to simulate the air quality from August 24 to September 06, 2016. It turns out the WRF-Chem model generally performed well when compared to the observed data from a large number of air quality monitoring sites and satellite measurements. During the period, an occurrence of stratospheric intrusion was observed, associated with tropopause fold and curved upper-level jet in East Asia. A fairly large number of stratospheric passive tracers that had quite positive correlation with O3 concentrations were also found over the southeast China. Besides, observed ground surface anomaly of high O3 and the accompanied low CO as well as humidity further implied the downward transport of O3 from the stratosphere. These results suggest that stratospheric ozone intrusion acts as an additional source of the near surface tropospheric ozone concentration, which deteriorates the O3 pollution in China. This helps to explain the difficulty of O3 control during the campaign period of the G20, and provides some insights into stratosphere-to-troposphere transport of O3.