Reflected Signal Research Articles

Archaeological investigation of burials preluded by ground penetrating radar and geospatial technologies

Ground penetrating radar (GPR) is a non-invasive and efficient scientific tool in burial analysis that can 'see the unseen,' answering both simple questions such as the existence and boundaries of burials, as well as more difficult questions like whether burials are intact or decayed. This paper reviews common reflection signatures associated with burials and applies the three fundamental GPR principles (dielectric contrast, scattering and polarity of reflections) to two distinct caseworks involving historical graves and civilian burials, which serve as benchmarks with known ground truth. It encompasses a third test case involving family tree research in a civilian cemetery, drawing upon the benchmarked results obtained from the first two caseworks. In adherence to geophysical signal reflection principles, our study discerns distinct hyperbolic traits associated with three burial types including intact shroud-wrapped, coffin burials, and decayed or mass-grave burials. A GPR-geospatial integration workflow incorporating GPR, aerial photogrammetry and global navigation satellite system - real time kinetics (GNSS-RTK), is derived to enhance the identification and investigation of burials using GPR. Our workflow encompasses a range of indicators for survey methods and burial classification, presenting a general framework for the systematic contextualization of tailored workflows to individual contexts. This work exemplifies the efficacy of GPR in the detection of burials that have been undisturbed for over a century in the soils of Hong Kong and how geophysics and geospatial science can address the limitations inherent in conventional desktop-based archaeological investigation. Its implications extend to professionals in diverse fields including historians, archaeologists, cemetery management officials, and even family members searching for their lost loved ones.

Carrier phase error overbounding using Gaussian mixture model for high-precision air navigation integrity monitoring

Carrier phase measurements play a pivotal role in enabling high-precision positioning within future air traffic systems and autonomous vehicles. Current aviation navigation integrity monitoring, primarily using pseudorange measurements, cannot meet system positioning accuracy and integrity requirements, leading to limitations and false alarm risks, especially in complex urban canyons with signal reflections. This work presents a carrier phase-based positioning integrity monitoring algorithm, utilizing the expectation–maximization algorithm to compute parameters for the Gaussian mixture model (GMM) and nonparametric bootstrapping to construct an overbounding model of carrier phase residuals. It establishes a precise protection level (PL) calculation method, which is validated with static and flight data, demonstrating accurate overbounding of carrier phase residuals using GMM. Protection levels during testing remain below the alarm limit and exceed system positioning error, ensuring normal system operation without false or missed alarms, confirming the effectiveness of the established carrier phase error overbounding model.

Water isotope ratios reflect convection intensity rather than rain type proportions in the pantropics.

Against the traditional view, a recently published theory argued that isotope ratios are higher in convective precipitation but lower in stratiform precipitation and proposed that isotope ratios reflect rain type proportions. This theory has been widely cited despite some early reservations. Whether the theory represents a faithful reflection of signals of water isotope ratios remains unclear. Here, we reassess its validity from different timescales and broader observations from the pantropics. Unexpectedly, our findings contradict the theory on daily, monthly, and even annual timescales. Pantropical precipitation isotope ratios remain strongly correlated to convection intensity but are independent of rain type proportions because stratiform precipitation isotope ratios cover a large range of values. We find that the theory has many serious weaknesses related to preferential data selection and suggest that new theories need to be validated at more locations on different timescales before gaining widespread acceptance.

RDA-Net: A multi-cascade network for DAS background noise attenuation

The increasing attention to seismic data collection under complex geological conditions is attributed to the utilization of distributed acoustic sensing (DAS), which has gained significant prominence due to the rapid development of exploration techniques. However, the presence of background noise in DAS has adverse effects on the quality of DAS data and significantly affects the accuracy of the subsequent migration and imaging processing. In addition, these interferences have a different generation mechanism from random noise in geophone-acquired seismic data, resulting in the degradation of traditional denoising techniques. Recent advancements have introduced convolutional neural networks (CNNs) for DAS background noise attenuation, showcasing their superiority over conventional techniques in terms of denoising capability and processing accuracy. Unfortunately, most CNN-based methods only have a single-scale architecture and cannot precisely accomplish denoising tasks when confronted with weak reflection signals in DAS records. In this research, we introduce a residual dense-connection attention network called RDA-Net to further augment the denoising ability. More precisely, the multi-scale structure in the backbone of RDA-Net is employed to capture informative features from the DAS data. Simultaneously, we employ a self-enhanced spatial attention module to refine and strengthen the primary features, thereby positively impacting feature representation and denoising performance. Furthermore, dense connections are also leveraged to reinforce the feature-interaction capability and effective features. Compared with typical conventional methods and CNN-based frameworks, RDA-Net can improve the signal-to-noise ratio over 24 dB and show superiority in weak signal recovery and intense DAS noise attenuation, both for synthetic and field records.

Development and Validation of a New Type of Displacement-Based Miniatured Laser Vibrometers.

Developing a miniatured laser vibrometer becomes important for many engineering areas, such as experimental and operational modal analyses, model validation, and structural health monitoring. Due to its compact size and light weight, a miniatured laser vibrometer can be attached to various mobilized platforms, such as an unmanned aerial vehicle and a robotic arm whose payloads can usually not be large, to achieve a flexible vibration measurement capability. However, integrating optics into a miniaturized laser vibrometer presents several challenges. These include signal interference from ghost reflectance signals generated by the sub-components of integrated photonics, polarization effects caused by waveguide structures, wavelength drifting due to the semiconductor laser, and the poorer noise characteristics of an integrated laser chip compared to a non-integrated circuit. This work proposes a novel chip-based high-precision laser vibrometer by incorporating two or more sets of quadrature demodulation networks into its design. An additional set of quadrature demodulation networks with a distinct reference arm delay line length can be used to conduct real-time compensation to mitigate linear interference caused by temperature and environmental variations. A series of vibration measurements with frequencies ranging from 0.1 Hz to 1 MHz were conducted using the proposed laser vibrometer to show its repeatability and accuracy in vibration and ultrasonic vibration measurements, and its robustness to test surface conditions. The proposed laser vibrometer has the advantage of directly measuring the displacement response of a vibrating structure rather than integrating its velocity response to yield the measured displacement with a conventional laser Doppler vibrometer.

Signal integrality analysis of sealed MRPC for muography

The signal quality is a crucial index to determine the performance of MRPC detector. This paper has analyzed the signal integrality of the sealed MRPC structure in detail with CST simulation. The simulation work includes the reflection of high frequency signal in readout line, the reflection of signal through via hole, the transmission of signals under multiplexing. The via hole will bring a reflection and overshoot to signal. The interconnection of the readout strips will lead to a reduction of the amplitude of the acquired signal and the jitter of the baseline is increased. A multiplexer will affect the quality of the signal to some extent, and this influence needs to be considered in the detector design.

Baseline-free damage localization in multilayer metallic spherical shell structures using guided wave tomography

Multi-layer metal spherical shell structure is widely used as the core component of pressure-bearing equipment, deep-sea exploration equipment and other critical areas due to its plane stress uniformity and high specific strength. Its long-term service in complex and harsh environments will inevitably produce a variety of defects and damages, which will affect the safety of the equipment in service. Ultrasonic guided wave inspection is a potential non-destructive testing method, but the multilayer metal bonding structure between the metal and non-metal bonding layer impedance difference is large, resulting in defects located in the internal reflection signal is difficult to propagate to the outer layer with a sensor, the multilayer spherical shell structure itself leads to the complexity of the guided wave propagation characteristics, it is difficult to extract the individual modes of the time information for the defects of the detection and localization. Therefore, in this paper, we propose a probabilistic damage existence imaging method for spherical shell structures and combine it with the virtual time reversal technique to detect and localize the defects on the inner and outer surfaces of multilayered metallic spherical shell structures without benchmarks. The results show that the newly proposed method can realize the accurate localization imaging of internal/external defects of multilayer metal spherical shell structures.

Frequency‐Locked Wireless Multifunctional Surface Acoustic Wave Sensors

AbstractSurface acoustic waves (SAWs) have shown great potential for developing sensors for structural health monitoring (SHM) and lab‐on‐a‐chip (LOC) applications. Existing SAW sensors mainly rely on measuring the frequency shifts of high‐frequency (e.g., >0.1 GHz) resonance peaks. This study presents frequency‐locked wireless multifunctional SAW sensors that enable multiple wireless sensing functions, including strain sensing, temperature measurement, water presence detection, and vibration sensing. These sensors leverage SAW resonators on piezoelectric chips, inductive coupling‐based wireless power transmission, and, particularly, a frequency‐locked wireless sensing mechanism that works at low frequencies (e.g., <0.1 GHz). This mechanism locks the input frequency on the slope of a sensor's reflection spectrum and monitors the reflection signal's amplitude change induced by the changes of sensing parameters. The proof‐of‐concept experiments show that these wireless sensors can operate in a low‐power active mode for on‐demand wireless strain measurement, temperature sensing, and water presence detection. Moreover, these sensors can operate in a power‐free passive mode for vibration sensing, with results that agree well with laser vibrometer measurements. It is anticipated that the designs and mechanisms of the frequency‐locked wireless SAW sensors will inspire researchers to develop future wireless multifunctional sensors for SHM and LOC applications.

Enhanced sensitivity of magnetic field sensing using L-shaped dual Terfenol-D-FBGs based on the Vernier effect and a dual-loop optoelectronic oscillator

We proposed and successfully validated improved the sensitivity of magnetic field sensing using the Vernier effect in a dual-loop optoelectronic oscillator (OEO) incorporating cascaded L-shaped Terfenol-D-FBGs. Thanks to the time delay introduced by the dispersion compensating fiber (DCF) in the OEO cavity, precise conversion between the sensing FBG wavelength changes and the OEO oscillation frequency can be achieved. The central wavelengths of the sensing Terfenol-D-FBG change along with the magnetic field. Therefore, the magnetic field can be monitored by measuring the oscillating frequency of the OEO. In the proposed scheme, two reflection signals from the cascaded FBGs pass through two fiber paths with slightly different lengths, inducing the Vernier effect in the OEO, which significantly enhances the sensitivity of the sensor. Additionally, a bias magnetic field is applied to the sensing Terfenol-D alloy to further improve the sensitivity. Experimental results demonstrate that by applying a bias magnetic field on both sides of the sensing Terfenol-D, the sensitivity of the single-loop OEO is increased from 421 Hz/mT to 560 Hz/mT. Utilizing the Vernier effect, the magnetic field sensitivity of the dual-loop OEO reaches up to −16.54 kHz/mT, with a 29-times improvement over the single-loop OEO's 560 Hz/mT. By finely adjusting the path length difference, we further enhance the performance of the sensor, achieving sensitivities of −24.58 kHz/mT, −28.49 kHz/mT and −35.94 kHz/mT, respectively, which are 44 times, 51 times and 64 times higher than the single-loop OEO structure. Besides, to overcome the influence of temperature, a temperature compensation sensor is designed using a pair of L-shaped Terfenol-D alloys to host the cascaded sensing FBG and reference FBG. The experimental results show that the proposed L-shaped Terfenol-D with cascaded FBG can reduce the magnetic field error caused by temperature crosstalk to 0.03 mT.

Analysis of 3D positioning error for multipath indoor VLC system

SummaryA comparative analysis of 3D positioning error for two different configurations using different layouts of visible light communication (VLC) systems is presented in this paper. The Received Signal Strength (RSS) has been implemented for indoor localization systems using Line‐of‐Sight (LoS) and diffused reflection signals. The room size for configuration‐1 is 5 m × 5 m × 3 m, and the distance between adjacent LEDs is 2.5 m, 2.00 m, and 1.5 m for cases‐1, case‐2, and case‐3, respectively, whereas the room size for configuration‐2 is 7 m × 7 m × 5 m, and the separation between the LEDs is 3.5 m, 3 m, and 2.5 m for their respective cases. Through investigation, it has been shown that when only LS signal is considered, the separation between LEDs may not be an issue because positioning error changes by a very small amount as the separation between LEDs changes. The results show that as the distance between adjacent LEDs decreases, the received signal strength for LoS and L‐R1 signals increases. However, positioning error and BER rise, while the bit rate falls. Furthermore, the positioning error Vs receiver plane height for all three cases in configuration‐1 is the same up to a height of 2.89 m, whereas the positioning error in configuration‐2 is the same up to 4.4 m for all cases. The positioning error for case‐1 decreases as the height in configuration‐1 exceeds 2.89 m. Similarly, after reaching a height of 4.4 m for case‐2, the positioning error in configuration‐2 decreases. The LoS positioning error versus semi angle of the LED as well as the FOV of the receiver has been simulated for different positions of the receiver in configuration‐1. The investigation shows that the minimum positioning error is achieved at and FOV equal to 66.660 for all the positions of the receiver in the room. Thus, before configuring a practical indoor VLC geometrical model, proper VLC configurations such as LED separation, FOV of the receiver, semi angle of LED, and receiver height should be chosen based on the room dimensions.

A transcranial multiple waves suppression method for plane wave imaging based on Radon transform

Transcranial ultrasound imaging presents a significant challenge due to the intricate interplay between ultrasound waves and the heterogeneous human skull. The skull’s presence induces distortion, refraction, multiple scattering, and reflection of ultrasound signals, thereby complicating the acquisition of high-quality images. Extracting reflections from the entire waveform is crucial yet exceedingly challenging, as intracranial reflections are often obscured by strong amplitude direct waves and multiple scattering. In this paper, a multiple wave suppression method for ultrasound plane wave imaging is proposed to mitigate the impact of skull interference. Drawing upon prior research, we developed an enhanced high-resolution linear Radon transform using the maximum entropy principle and Bayesian method, facilitating wavefield separation. We detailed the process of wave field separation in the Radon domain through simulation of a model with a high velocity layer. When plane waves emitted at any steering angles, both multiple waves and first arrival waves manifested as distinct energy points. In the brain simulation, we contrasted the characteristic differences between skull reflection and brain-internal signal in Radon domain, and demonstrated that multiples suppression method reduces side and grating lobe levels by approximately 30 dB. Finally, we executed in vitro experiments using a monkey skull to separate weak intracranial reflection signals from strong skull reflections, enhancing the contrast-to-noise ratio by 85 % compared to conventional method using full waveform. This study deeply explores the effect of multiples on effective signal separation, addresses the complexity of wavefield separation, and verifies its efficacy through imaging, thereby significantly advancing ultrasound transcranial imaging techniques.

Tightly coupled integration of vector HD map, LiDAR, GNSS, and INS for precise vehicle navigation in GNSS-challenging environment

ABSTRACT High-precision positioning and navigation are necessary for autonomous driving. GNSS RTK and INS integrated system is commonly used in vehicular navigation. But it suffers from severe signal reflections and blockages of GNSS signals and error accumulation of INS with MEMS-IMU in GNSS-challenging environment. In addition, high-definition (HD) maps in vector format and light detection and ranging (LiDAR) are two common options for intelligent vehicles. A tightly coupled localization method with vector HD map, LiDAR, GNSS RTK, and INS is proposed to take advantage of their complementary characteristics to accurately navigate a vehicle in a GNSS-challenging environment. The method is based on a particle filter (PF) framework. Lateral positions of particles are constrained by LiDAR measurements and lane information in the vector HD map. A constrained and damped LAMBDA searching method is proposed to update particle weights, aiming to find accurate longitudinal localization. Experimental results prove that our method can maintain sub-meter level horizontal positioning accuracy in GNSS-challenging environment, with improvements of (77%, 75%, 64% and 65%), regarding to three-dimensional position and yaw, compared to traditional GNSS-RTK/INS integration, while the improvement of the popular GNSS-RTK/INS/LiDAR integrated framework LIO-SAM is (16%, 53%, 48% and 51%).

Enhancing Radar Image Classification with Autoencoder-CNN Hybrid System

The tracking, analysis, and classification of human movements can be crucial, particularly in areas such as elderly care, healthcare, and infant care. Typically, such tracking is done remotely with cameras. However, radar systems have emerged as significant methods and tools for these tasks due to their advantages such as privacy, wireless operation, and the ability to work through walls. By converting reflected radar signals from targets into images, human activities can be classified using powerful classification tools like deep learning. In this study, range-Doppler images of behind-the-wall human movements obtained with a radar system consisting of one transmitter and four receiver antennas were classified. Since the data collected from the four receiver antennas are in different positions, the collected reflection signals also differ. The signals collected with the range-time matrix content were divided into positive and negative parts, resulting in eight images from the four antennas. Instead of using all the data in CNN training, the images were first subjected to a reconstruction process with an autoencoder to reduce differences. As a result, it was observed that reconstructing the images with an autoencoder before classification with CNN increased the classification success. In conclusion, it was observed that the classification success of radar images can be increased by using a hybrid system with an autoencoder to reconstruct the images before classification with CNN.

Discovering Millimeter Wave Technology forNext-Generation (5g) Communication: A Review

Millimeter Wave (mmWave) frequencies, a factor of the electromagnetic spectrum, provide a varied array of frequency bands that can be connected for the development of 5G mobile communication technologies. These frequency bands proposal a substantial view for advancing data transfer rates and network capacity inside the area of wireless communication. However, the employment of mmWave technology faces numerous challenges and obstacles that must be fixed. One of the key challenges relates to propagation losses, which have the possible to disturb the effectiveness of signal propagation over unlimited distances. Furthermore, accomplishing accurate beam alignment represents a vital obstacle in guaranteeing precise signal transmission between transmitters and receivers, thus optimizing the total efficiency of 5G infrastructures. An additional difficulty to be overcame is the existence of signal reflection paths, which may result in interference and signal decline if not effectively managed. In conflict, technique similar to beamforming policies is vital progression revealing the potential of mmWave which is utilize many antennas operating together and automatically instruction and intensive the signal to the selected recipient with notable precision. In addition,participating phased antenna arrays as improvements in antenna technology and reduced antennas and integrated circuits is extra advancement in 5G technology. Furthermore, scattering applications challenging high precision as a result of the high-frequency nature of mmWave and realizing multi-gigabit-per-second has the beneficial comprise the possible for maximum-capacity wireless broadcast at high data rates. The determination of these challenges is authoritative for revealing the comprehensive potential of mmWave technology and exploiting on the advantages it can offer to the realmfield of mobile communications.

Correction of microwave pulse reflection by digital filters in superconducting quantum circuits

Abstract Reducing the control error is vital for high-fidelity digital and analog quantum operations. In superconducting circuits, one disagreeable error arises from the reflection of microwave signals due to impedance mismatch in the control chain. Here, we demonstrate a reflection cancelation method when considering that there are two reflection nodes on the control line. We propose to generate the pre-distortion pulse by passing the envelopes of the microwave signal through digital filters, which enables real-time reflection correction when integrated into the field-programmable gate array (FPGA). We achieve a reduction of single-qubit gate infidelity from 0.67% to 0.11% after eliminating microwave reflection. Real-time correction of microwave reflection paves the way for precise control and manipulation of the qubit state and would ultimately enhance the performance of algorithms and simulations executed on quantum processors.

Optimizing the Deployment of an Aerial Base Station and the Phase-Shift of a Ground Reconfigurable Intelligent Surface for Wireless Communication Systems Using Deep Reinforcement Learning

In wireless networks, drone base stations (DBSs) offer significant benefits in terms of Quality of Service (QoS) improvement due to their line-of-sight (LoS) transmission capabilities and adaptability. However, LoS links can suffer degradation in complex propagation environments, especially in urban areas with dense structures like buildings. As a promising technology to enhance the wireless communication networks, reconfigurable intelligent surfaces (RIS) have emerged in various Internet of Things (IoT) applications by adjusting the amplitude and phase of reflected signals, thereby improving signal strength and network efficiency. This study aims to propose a novel approach to enhance communication coverage and throughput for mobile ground users by intelligently leveraging signal reflection from DBSs using ground-based RIS. We employ Deep Reinforcement Learning (DRL) to optimize both the DBS location and RIS phase-shifts. Numerical results demonstrate significant improvements in system performance, including communication quality and network throughput, validating the effectiveness of the proposed approach.

Moho Imaging with Fiber Borehole Strainmeters Based on Ambient Noise Autocorrelation.

Moho tomography is important for studying the deep Earth structure and geodynamics, and fiber borehole strainmeters are broadband, low-noise, and attractive tools for seismic observation. Recently, many studies have shown that fiber optic seismic sensors can be used for subsurface structure imaging based on ambient noise cross-correlation, similar to conventional geophones. However, this array-dependent cross-correlation method is not suitable for fiber borehole strainmeters. Here, we developed a Moho imaging scheme for the characteristics of fiber borehole strainmeters based on ambient noise autocorrelation. S-wave reflection signals were extracted from the ambient noise through a series of processing steps, including phase autocorrelation (PAC), phase-weighted stacking (PWS), etc. Subsequently, the time-to-depth conversion crustal thickness beneath the station was calculated. We applied our scheme to continuous four-component recordings from four fiber borehole strainmeters in Lu'an, Anhui Province, China. The obtained Moho depth was consistent with the previous research results. Our work shows that this method is suitable for Moho imaging with fiber borehole strainmeters without relying on the number of stations.

A solid-state sensor based on nanosilica/methylcellulose composite and chromoionophore 8-carboxamidoquinoline derivative as an optode pellet for Zinc(II) ion

Zinc is known to play a central role in the immune system that creates a resistance barrier against viral infection. A naked-eye zinc(II) (Zn2+) ion sensor of 2-oxo-2-(quinolin-8-ylamino)acetic acid (OQAA) as a recognition probe was designed and synthesized. By inserting functional groups of amide and carboxylic acid into a conjugated molecule of 8-aminoquinoline, Zn2+ ion (borderline acid) coordinated favorably with aromatic nitrogen atoms (borderline bases) and N-amide. Characterization of OQAA was done by FTIR, 1H NMR, 13C NMR, UV–vis, and ESI-MS. The Zn2+ ion chemosensor was fabricated by immobilizing OQAA chromogenic ionophore onto a nanosilica/methylcellulose (Si/MC) composite pellet to form a solid-state pellet sensor. A distinguishable color change from pale light-yellow to yellowish-brown of the optode pellet was observed in the presence of Zn2+ ions. Optimization of the pellet-based Zn2+ ion sensor was carried out with a fiber optic reflectance spectrophotometer at 0.25 M OQAA in tris buffer (pH 7.4). The optical fiber sensor demonstrated a linear reflectance response between 0.8 M and 2.4 M Zn2+ ion (R²=0.9741) at λmax=635 nm with a limit of detection (LOD) of 3.6 × 10−2 M and response time of 5 min. The reflectance sensor exhibited good selectivity towards Zn2+ ion over other competitive heavy metals e.g., iron(II) (Fe2+), nickel(II) (Ni2+), cobalt(II) (Co2+), copper(II)(Cu2+) and cadmium(Cd2+) ions. Additionally, the proposed Zn2+ ion sensor showed high shelf-life stability of 75 days operational duration with 80.0 % initial optical response remaining. The quick, easy, and uncomplicated preparation of the OQAA-modified Si/MC (OQAA-Si/MC) composite pellet novel material gave a reproducible reflectance signal towards the determination of Zn2+ ion concentration with a low relative standard deviation (RSD) obtained at ∼1.0 % (n=12), and was reusable three times for assay of Zn2+ ion (RSD=1.3 %). The optical chemosensor imparts remarkable color change towards the analysis of Zn2+ ions, and could offer direct and fast in-situ visual inspection of Zn2+ ion levels in complex samples without an instrument.

Advanced impedance mismatch technique for detecting faults in photovoltaic systems

AbstractThis paper presents a comprehensive exploration of the Advanced Impedance Mismatch Technique (AIMT), a novel approach designed for the accurate detection of simultaneous and varied faults within photovoltaic (PV) systems. This investigation integrates a spectrum of fault detection strategies, pinpointing reflectometry as a notably effective tool. Despite its utility, conventional reflectometry applications face critical constraints, notably the limitation to identify only the primary fault location within a PV array and the inability to distinguish between different fault types. This work introduces an innovative mathematical model that estimates the impedance of PV modules, enhancing the reflectometry method to enable the precise identification and localization of multiple defective modules within a string. The proposed technique exhibits a remarkable sensitivity to detect slight impedance differences between a functional PV string and one with defective modules. The validity of the AIMT's mathematical model is corroborated through simulation experiments on a string of seven PV modules afflicted with multiple simultaneous faults. These experiments rigorously evaluate the technique's accuracy in pinpointing the locations of defective modules within a PV string. The outcomes of our proposed ‐times faster AIMT reveal a strong concordance between the simulated reflective signals and the impedance values forecasted by the model, highlighting the proposed method's proficiency in the detailed detection and diagnosis of progressive faults within PV systems.

Time-Series Data Augmentation for Improving Multi-Class Classification Performance

This paper proposes a new approach to classify and evaluate defects in concrete structures automatically. To overcome the limitations of defect detection methods that traditionally relied on expert visual observation, the reflection signal of electromagnetic pulses is extracted as time-series data and used to analyze the propagation characteristics of each defect. This study uses deep learning models to analyze these time-series data and classify defects. Since anomaly detection data has more normal data than anomaly data, data augmentation methods such as Time Warping, Noise Injection, Smoothing, Trend Shifting, etc., were applied to solve the problem of data imbalance and overfitting. Among them, Noise Injection showed the best performance. The generalization performance of the proposed method was evaluated through performance evaluation using LSTM, GRU, and TCN models, and LSTM models showed the highest performance. The study results show that the proposed method effectively classifies defect types in concrete structures and can solve the limitations of existing methods by automatic classification through deep learning models. In addition, it was confirmed that the model's performance could be improved by improving the amount and diversity of data by selecting and applying appropriate data augmentation methods. The contribution of the research is to present a new approach that automates the defect detection and classification task of concrete structures and provides high accuracy and efficiency.